[32mINFO    [0m [32m2024-09-05 10:01:31,408 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-05 10:01:31,410 | [34mCommand: main.py --local-rank=0 --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --pretrain_model_path /home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth --options text_encoder_type=/home/jiask/Open-GroundingDino-main/weights/bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-05 10:01:31,411 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-05 10:01:31,413 | [34mworld size: 9[0m
[32mINFO    [0m [32m2024-09-05 10:01:31,414 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:01:31,415 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:01:31,416 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='config/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_backend='nccl', dist_url='env://', distributed=True, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, gpu=0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': '/home/jiask/Open-GroundingDino-main/weights/bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path='/home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth', query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='/home/jiask/Open-GroundingDino-main/weights/bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=9)
[0m
[36mDEBUG   [0m [36m2024-09-05 10:01:31,418 | [34mbuild model ... ...[0m
[32mINFO    [0m [32m2024-09-05 10:26:23,532 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-05 10:26:23,533 | [34mCommand: main.py --local-rank=0 --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --pretrain_model_path /home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-05 10:26:23,535 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-05 10:27:12,595 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-05 10:27:12,596 | [34mCommand: main.py --local-rank=0 --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --pretrain_model_path /home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-05 10:27:12,598 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-05 10:27:12,600 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-05 10:27:12,600 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:27:12,601 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:27:12,602 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='config/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_backend='nccl', dist_url='env://', distributed=True, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, gpu=0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path='/home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth', query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-05 10:27:12,604 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-05 10:27:24,887 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-05 10:27:25,027 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-05 10:27:25,040 | [34mparams before freezing:
{
  "module.transformer.level_embed": 1024,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.0.norm1.weight": 256,
  "module.transformer.encoder.layers.0.norm1.bias": 256,
  "module.transformer.encoder.layers.0.linear1.weight": 524288,
  "module.transformer.encoder.layers.0.linear1.bias": 2048,
  "module.transformer.encoder.layers.0.linear2.weight": 524288,
  "module.transformer.encoder.layers.0.linear2.bias": 256,
  "module.transformer.encoder.layers.0.norm2.weight": 256,
  "module.transformer.encoder.layers.0.norm2.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.1.norm1.weight": 256,
  "module.transformer.encoder.layers.1.norm1.bias": 256,
  "module.transformer.encoder.layers.1.linear1.weight": 524288,
  "module.transformer.encoder.layers.1.linear1.bias": 2048,
  "module.transformer.encoder.layers.1.linear2.weight": 524288,
  "module.transformer.encoder.layers.1.linear2.bias": 256,
  "module.transformer.encoder.layers.1.norm2.weight": 256,
  "module.transformer.encoder.layers.1.norm2.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.2.norm1.weight": 256,
  "module.transformer.encoder.layers.2.norm1.bias": 256,
  "module.transformer.encoder.layers.2.linear1.weight": 524288,
  "module.transformer.encoder.layers.2.linear1.bias": 2048,
  "module.transformer.encoder.layers.2.linear2.weight": 524288,
  "module.transformer.encoder.layers.2.linear2.bias": 256,
  "module.transformer.encoder.layers.2.norm2.weight": 256,
  "module.transformer.encoder.layers.2.norm2.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.3.norm1.weight": 256,
  "module.transformer.encoder.layers.3.norm1.bias": 256,
  "module.transformer.encoder.layers.3.linear1.weight": 524288,
  "module.transformer.encoder.layers.3.linear1.bias": 2048,
  "module.transformer.encoder.layers.3.linear2.weight": 524288,
  "module.transformer.encoder.layers.3.linear2.bias": 256,
  "module.transformer.encoder.layers.3.norm2.weight": 256,
  "module.transformer.encoder.layers.3.norm2.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.4.norm1.weight": 256,
  "module.transformer.encoder.layers.4.norm1.bias": 256,
  "module.transformer.encoder.layers.4.linear1.weight": 524288,
  "module.transformer.encoder.layers.4.linear1.bias": 2048,
  "module.transformer.encoder.layers.4.linear2.weight": 524288,
  "module.transformer.encoder.layers.4.linear2.bias": 256,
  "module.transformer.encoder.layers.4.norm2.weight": 256,
  "module.transformer.encoder.layers.4.norm2.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.5.norm1.weight": 256,
  "module.transformer.encoder.layers.5.norm1.bias": 256,
  "module.transformer.encoder.layers.5.linear1.weight": 524288,
  "module.transformer.encoder.layers.5.linear1.bias": 2048,
  "module.transformer.encoder.layers.5.linear2.weight": 524288,
  "module.transformer.encoder.layers.5.linear2.bias": 256,
  "module.transformer.encoder.layers.5.norm2.weight": 256,
  "module.transformer.encoder.layers.5.norm2.bias": 256,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.0.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.0.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear2.bias": 256,
  "module.transformer.encoder.text_layers.0.norm1.weight": 256,
  "module.transformer.encoder.text_layers.0.norm1.bias": 256,
  "module.transformer.encoder.text_layers.0.norm2.weight": 256,
  "module.transformer.encoder.text_layers.0.norm2.bias": 256,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.1.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.1.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear2.bias": 256,
  "module.transformer.encoder.text_layers.1.norm1.weight": 256,
  "module.transformer.encoder.text_layers.1.norm1.bias": 256,
  "module.transformer.encoder.text_layers.1.norm2.weight": 256,
  "module.transformer.encoder.text_layers.1.norm2.bias": 256,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.2.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.2.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear2.bias": 256,
  "module.transformer.encoder.text_layers.2.norm1.weight": 256,
  "module.transformer.encoder.text_layers.2.norm1.bias": 256,
  "module.transformer.encoder.text_layers.2.norm2.weight": 256,
  "module.transformer.encoder.text_layers.2.norm2.bias": 256,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.3.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.3.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear2.bias": 256,
  "module.transformer.encoder.text_layers.3.norm1.weight": 256,
  "module.transformer.encoder.text_layers.3.norm1.bias": 256,
  "module.transformer.encoder.text_layers.3.norm2.weight": 256,
  "module.transformer.encoder.text_layers.3.norm2.bias": 256,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.4.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.4.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear2.bias": 256,
  "module.transformer.encoder.text_layers.4.norm1.weight": 256,
  "module.transformer.encoder.text_layers.4.norm1.bias": 256,
  "module.transformer.encoder.text_layers.4.norm2.weight": 256,
  "module.transformer.encoder.text_layers.4.norm2.bias": 256,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.5.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.5.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear2.bias": 256,
  "module.transformer.encoder.text_layers.5.norm1.weight": 256,
  "module.transformer.encoder.text_layers.5.norm1.bias": 256,
  "module.transformer.encoder.text_layers.5.norm2.weight": 256,
  "module.transformer.encoder.text_layers.5.norm2.bias": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm1.weight": 256,
  "module.transformer.decoder.layers.0.norm1.bias": 256,
  "module.transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.catext_norm.weight": 256,
  "module.transformer.decoder.layers.0.catext_norm.bias": 256,
  "module.transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm2.weight": 256,
  "module.transformer.decoder.layers.0.norm2.bias": 256,
  "module.transformer.decoder.layers.0.linear1.weight": 524288,
  "module.transformer.decoder.layers.0.linear1.bias": 2048,
  "module.transformer.decoder.layers.0.linear2.weight": 524288,
  "module.transformer.decoder.layers.0.linear2.bias": 256,
  "module.transformer.decoder.layers.0.norm3.weight": 256,
  "module.transformer.decoder.layers.0.norm3.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm1.weight": 256,
  "module.transformer.decoder.layers.1.norm1.bias": 256,
  "module.transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.catext_norm.weight": 256,
  "module.transformer.decoder.layers.1.catext_norm.bias": 256,
  "module.transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm2.weight": 256,
  "module.transformer.decoder.layers.1.norm2.bias": 256,
  "module.transformer.decoder.layers.1.linear1.weight": 524288,
  "module.transformer.decoder.layers.1.linear1.bias": 2048,
  "module.transformer.decoder.layers.1.linear2.weight": 524288,
  "module.transformer.decoder.layers.1.linear2.bias": 256,
  "module.transformer.decoder.layers.1.norm3.weight": 256,
  "module.transformer.decoder.layers.1.norm3.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm1.weight": 256,
  "module.transformer.decoder.layers.2.norm1.bias": 256,
  "module.transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.catext_norm.weight": 256,
  "module.transformer.decoder.layers.2.catext_norm.bias": 256,
  "module.transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm2.weight": 256,
  "module.transformer.decoder.layers.2.norm2.bias": 256,
  "module.transformer.decoder.layers.2.linear1.weight": 524288,
  "module.transformer.decoder.layers.2.linear1.bias": 2048,
  "module.transformer.decoder.layers.2.linear2.weight": 524288,
  "module.transformer.decoder.layers.2.linear2.bias": 256,
  "module.transformer.decoder.layers.2.norm3.weight": 256,
  "module.transformer.decoder.layers.2.norm3.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm1.weight": 256,
  "module.transformer.decoder.layers.3.norm1.bias": 256,
  "module.transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.catext_norm.weight": 256,
  "module.transformer.decoder.layers.3.catext_norm.bias": 256,
  "module.transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm2.weight": 256,
  "module.transformer.decoder.layers.3.norm2.bias": 256,
  "module.transformer.decoder.layers.3.linear1.weight": 524288,
  "module.transformer.decoder.layers.3.linear1.bias": 2048,
  "module.transformer.decoder.layers.3.linear2.weight": 524288,
  "module.transformer.decoder.layers.3.linear2.bias": 256,
  "module.transformer.decoder.layers.3.norm3.weight": 256,
  "module.transformer.decoder.layers.3.norm3.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm1.weight": 256,
  "module.transformer.decoder.layers.4.norm1.bias": 256,
  "module.transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.catext_norm.weight": 256,
  "module.transformer.decoder.layers.4.catext_norm.bias": 256,
  "module.transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm2.weight": 256,
  "module.transformer.decoder.layers.4.norm2.bias": 256,
  "module.transformer.decoder.layers.4.linear1.weight": 524288,
  "module.transformer.decoder.layers.4.linear1.bias": 2048,
  "module.transformer.decoder.layers.4.linear2.weight": 524288,
  "module.transformer.decoder.layers.4.linear2.bias": 256,
  "module.transformer.decoder.layers.4.norm3.weight": 256,
  "module.transformer.decoder.layers.4.norm3.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm1.weight": 256,
  "module.transformer.decoder.layers.5.norm1.bias": 256,
  "module.transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.catext_norm.weight": 256,
  "module.transformer.decoder.layers.5.catext_norm.bias": 256,
  "module.transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm2.weight": 256,
  "module.transformer.decoder.layers.5.norm2.bias": 256,
  "module.transformer.decoder.layers.5.linear1.weight": 524288,
  "module.transformer.decoder.layers.5.linear1.bias": 2048,
  "module.transformer.decoder.layers.5.linear2.weight": 524288,
  "module.transformer.decoder.layers.5.linear2.bias": 256,
  "module.transformer.decoder.layers.5.norm3.weight": 256,
  "module.transformer.decoder.layers.5.norm3.bias": 256,
  "module.transformer.decoder.norm.weight": 256,
  "module.transformer.decoder.norm.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "module.transformer.decoder.ref_point_head.layers.0.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "module.transformer.decoder.ref_point_head.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "module.transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "module.transformer.tgt_embed.weight": 230400,
  "module.transformer.enc_output.weight": 65536,
  "module.transformer.enc_output.bias": 256,
  "module.transformer.enc_output_norm.weight": 256,
  "module.transformer.enc_output_norm.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "module.transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "module.bert.embeddings.word_embeddings.weight": 23440896,
  "module.bert.embeddings.position_embeddings.weight": 393216,
  "module.bert.embeddings.token_type_embeddings.weight": 1536,
  "module.bert.embeddings.LayerNorm.weight": 768,
  "module.bert.embeddings.LayerNorm.bias": 768,
  "module.bert.encoder.layer.0.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.query.bias": 768,
  "module.bert.encoder.layer.0.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.key.bias": 768,
  "module.bert.encoder.layer.0.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.value.bias": 768,
  "module.bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.0.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.0.output.dense.weight": 2359296,
  "module.bert.encoder.layer.0.output.dense.bias": 768,
  "module.bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.1.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.query.bias": 768,
  "module.bert.encoder.layer.1.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.key.bias": 768,
  "module.bert.encoder.layer.1.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.value.bias": 768,
  "module.bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.1.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.1.output.dense.weight": 2359296,
  "module.bert.encoder.layer.1.output.dense.bias": 768,
  "module.bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.2.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.query.bias": 768,
  "module.bert.encoder.layer.2.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.key.bias": 768,
  "module.bert.encoder.layer.2.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.value.bias": 768,
  "module.bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.2.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.2.output.dense.weight": 2359296,
  "module.bert.encoder.layer.2.output.dense.bias": 768,
  "module.bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.3.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.query.bias": 768,
  "module.bert.encoder.layer.3.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.key.bias": 768,
  "module.bert.encoder.layer.3.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.value.bias": 768,
  "module.bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.3.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.3.output.dense.weight": 2359296,
  "module.bert.encoder.layer.3.output.dense.bias": 768,
  "module.bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.4.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.query.bias": 768,
  "module.bert.encoder.layer.4.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.key.bias": 768,
  "module.bert.encoder.layer.4.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.value.bias": 768,
  "module.bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.4.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.4.output.dense.weight": 2359296,
  "module.bert.encoder.layer.4.output.dense.bias": 768,
  "module.bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.5.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.query.bias": 768,
  "module.bert.encoder.layer.5.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.key.bias": 768,
  "module.bert.encoder.layer.5.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.value.bias": 768,
  "module.bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.5.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.5.output.dense.weight": 2359296,
  "module.bert.encoder.layer.5.output.dense.bias": 768,
  "module.bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.6.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.query.bias": 768,
  "module.bert.encoder.layer.6.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.key.bias": 768,
  "module.bert.encoder.layer.6.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.value.bias": 768,
  "module.bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.6.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.6.output.dense.weight": 2359296,
  "module.bert.encoder.layer.6.output.dense.bias": 768,
  "module.bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.7.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.query.bias": 768,
  "module.bert.encoder.layer.7.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.key.bias": 768,
  "module.bert.encoder.layer.7.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.value.bias": 768,
  "module.bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.7.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.7.output.dense.weight": 2359296,
  "module.bert.encoder.layer.7.output.dense.bias": 768,
  "module.bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.8.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.query.bias": 768,
  "module.bert.encoder.layer.8.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.key.bias": 768,
  "module.bert.encoder.layer.8.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.value.bias": 768,
  "module.bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.8.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.8.output.dense.weight": 2359296,
  "module.bert.encoder.layer.8.output.dense.bias": 768,
  "module.bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.9.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.query.bias": 768,
  "module.bert.encoder.layer.9.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.key.bias": 768,
  "module.bert.encoder.layer.9.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.value.bias": 768,
  "module.bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.9.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.9.output.dense.weight": 2359296,
  "module.bert.encoder.layer.9.output.dense.bias": 768,
  "module.bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.10.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.query.bias": 768,
  "module.bert.encoder.layer.10.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.key.bias": 768,
  "module.bert.encoder.layer.10.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.value.bias": 768,
  "module.bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.10.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.10.output.dense.weight": 2359296,
  "module.bert.encoder.layer.10.output.dense.bias": 768,
  "module.bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.11.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.query.bias": 768,
  "module.bert.encoder.layer.11.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.key.bias": 768,
  "module.bert.encoder.layer.11.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.value.bias": 768,
  "module.bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.11.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.11.output.dense.weight": 2359296,
  "module.bert.encoder.layer.11.output.dense.bias": 768,
  "module.bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "module.feat_map.weight": 196608,
  "module.feat_map.bias": 256,
  "module.input_proj.0.0.weight": 49152,
  "module.input_proj.0.0.bias": 256,
  "module.input_proj.0.1.weight": 256,
  "module.input_proj.0.1.bias": 256,
  "module.input_proj.1.0.weight": 98304,
  "module.input_proj.1.0.bias": 256,
  "module.input_proj.1.1.weight": 256,
  "module.input_proj.1.1.bias": 256,
  "module.input_proj.2.0.weight": 196608,
  "module.input_proj.2.0.bias": 256,
  "module.input_proj.2.1.weight": 256,
  "module.input_proj.2.1.bias": 256,
  "module.input_proj.3.0.weight": 1769472,
  "module.input_proj.3.0.bias": 256,
  "module.input_proj.3.1.weight": 256,
  "module.input_proj.3.1.bias": 256,
  "module.backbone.0.patch_embed.proj.weight": 4608,
  "module.backbone.0.patch_embed.proj.bias": 96,
  "module.backbone.0.patch_embed.norm.weight": 96,
  "module.backbone.0.patch_embed.norm.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.downsample.reduction.weight": 73728,
  "module.backbone.0.layers.0.downsample.norm.weight": 384,
  "module.backbone.0.layers.0.downsample.norm.bias": 384,
  "module.backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.downsample.reduction.weight": 294912,
  "module.backbone.0.layers.1.downsample.norm.weight": 768,
  "module.backbone.0.layers.1.downsample.norm.bias": 768,
  "module.backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "module.backbone.0.layers.2.downsample.norm.weight": 1536,
  "module.backbone.0.layers.2.downsample.norm.bias": 1536,
  "module.backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "module.backbone.0.norm1.weight": 192,
  "module.backbone.0.norm1.bias": 192,
  "module.backbone.0.norm2.weight": 384,
  "module.backbone.0.norm2.bias": 384,
  "module.backbone.0.norm3.weight": 768,
  "module.backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-05 10:27:25,089 | [34mparams after freezing:
{
  "module.transformer.level_embed": 1024,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.0.norm1.weight": 256,
  "module.transformer.encoder.layers.0.norm1.bias": 256,
  "module.transformer.encoder.layers.0.linear1.weight": 524288,
  "module.transformer.encoder.layers.0.linear1.bias": 2048,
  "module.transformer.encoder.layers.0.linear2.weight": 524288,
  "module.transformer.encoder.layers.0.linear2.bias": 256,
  "module.transformer.encoder.layers.0.norm2.weight": 256,
  "module.transformer.encoder.layers.0.norm2.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.1.norm1.weight": 256,
  "module.transformer.encoder.layers.1.norm1.bias": 256,
  "module.transformer.encoder.layers.1.linear1.weight": 524288,
  "module.transformer.encoder.layers.1.linear1.bias": 2048,
  "module.transformer.encoder.layers.1.linear2.weight": 524288,
  "module.transformer.encoder.layers.1.linear2.bias": 256,
  "module.transformer.encoder.layers.1.norm2.weight": 256,
  "module.transformer.encoder.layers.1.norm2.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.2.norm1.weight": 256,
  "module.transformer.encoder.layers.2.norm1.bias": 256,
  "module.transformer.encoder.layers.2.linear1.weight": 524288,
  "module.transformer.encoder.layers.2.linear1.bias": 2048,
  "module.transformer.encoder.layers.2.linear2.weight": 524288,
  "module.transformer.encoder.layers.2.linear2.bias": 256,
  "module.transformer.encoder.layers.2.norm2.weight": 256,
  "module.transformer.encoder.layers.2.norm2.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.3.norm1.weight": 256,
  "module.transformer.encoder.layers.3.norm1.bias": 256,
  "module.transformer.encoder.layers.3.linear1.weight": 524288,
  "module.transformer.encoder.layers.3.linear1.bias": 2048,
  "module.transformer.encoder.layers.3.linear2.weight": 524288,
  "module.transformer.encoder.layers.3.linear2.bias": 256,
  "module.transformer.encoder.layers.3.norm2.weight": 256,
  "module.transformer.encoder.layers.3.norm2.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.4.norm1.weight": 256,
  "module.transformer.encoder.layers.4.norm1.bias": 256,
  "module.transformer.encoder.layers.4.linear1.weight": 524288,
  "module.transformer.encoder.layers.4.linear1.bias": 2048,
  "module.transformer.encoder.layers.4.linear2.weight": 524288,
  "module.transformer.encoder.layers.4.linear2.bias": 256,
  "module.transformer.encoder.layers.4.norm2.weight": 256,
  "module.transformer.encoder.layers.4.norm2.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.5.norm1.weight": 256,
  "module.transformer.encoder.layers.5.norm1.bias": 256,
  "module.transformer.encoder.layers.5.linear1.weight": 524288,
  "module.transformer.encoder.layers.5.linear1.bias": 2048,
  "module.transformer.encoder.layers.5.linear2.weight": 524288,
  "module.transformer.encoder.layers.5.linear2.bias": 256,
  "module.transformer.encoder.layers.5.norm2.weight": 256,
  "module.transformer.encoder.layers.5.norm2.bias": 256,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.0.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.0.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear2.bias": 256,
  "module.transformer.encoder.text_layers.0.norm1.weight": 256,
  "module.transformer.encoder.text_layers.0.norm1.bias": 256,
  "module.transformer.encoder.text_layers.0.norm2.weight": 256,
  "module.transformer.encoder.text_layers.0.norm2.bias": 256,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.1.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.1.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear2.bias": 256,
  "module.transformer.encoder.text_layers.1.norm1.weight": 256,
  "module.transformer.encoder.text_layers.1.norm1.bias": 256,
  "module.transformer.encoder.text_layers.1.norm2.weight": 256,
  "module.transformer.encoder.text_layers.1.norm2.bias": 256,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.2.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.2.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear2.bias": 256,
  "module.transformer.encoder.text_layers.2.norm1.weight": 256,
  "module.transformer.encoder.text_layers.2.norm1.bias": 256,
  "module.transformer.encoder.text_layers.2.norm2.weight": 256,
  "module.transformer.encoder.text_layers.2.norm2.bias": 256,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.3.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.3.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear2.bias": 256,
  "module.transformer.encoder.text_layers.3.norm1.weight": 256,
  "module.transformer.encoder.text_layers.3.norm1.bias": 256,
  "module.transformer.encoder.text_layers.3.norm2.weight": 256,
  "module.transformer.encoder.text_layers.3.norm2.bias": 256,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.4.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.4.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear2.bias": 256,
  "module.transformer.encoder.text_layers.4.norm1.weight": 256,
  "module.transformer.encoder.text_layers.4.norm1.bias": 256,
  "module.transformer.encoder.text_layers.4.norm2.weight": 256,
  "module.transformer.encoder.text_layers.4.norm2.bias": 256,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.5.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.5.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear2.bias": 256,
  "module.transformer.encoder.text_layers.5.norm1.weight": 256,
  "module.transformer.encoder.text_layers.5.norm1.bias": 256,
  "module.transformer.encoder.text_layers.5.norm2.weight": 256,
  "module.transformer.encoder.text_layers.5.norm2.bias": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm1.weight": 256,
  "module.transformer.decoder.layers.0.norm1.bias": 256,
  "module.transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.catext_norm.weight": 256,
  "module.transformer.decoder.layers.0.catext_norm.bias": 256,
  "module.transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm2.weight": 256,
  "module.transformer.decoder.layers.0.norm2.bias": 256,
  "module.transformer.decoder.layers.0.linear1.weight": 524288,
  "module.transformer.decoder.layers.0.linear1.bias": 2048,
  "module.transformer.decoder.layers.0.linear2.weight": 524288,
  "module.transformer.decoder.layers.0.linear2.bias": 256,
  "module.transformer.decoder.layers.0.norm3.weight": 256,
  "module.transformer.decoder.layers.0.norm3.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm1.weight": 256,
  "module.transformer.decoder.layers.1.norm1.bias": 256,
  "module.transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.catext_norm.weight": 256,
  "module.transformer.decoder.layers.1.catext_norm.bias": 256,
  "module.transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm2.weight": 256,
  "module.transformer.decoder.layers.1.norm2.bias": 256,
  "module.transformer.decoder.layers.1.linear1.weight": 524288,
  "module.transformer.decoder.layers.1.linear1.bias": 2048,
  "module.transformer.decoder.layers.1.linear2.weight": 524288,
  "module.transformer.decoder.layers.1.linear2.bias": 256,
  "module.transformer.decoder.layers.1.norm3.weight": 256,
  "module.transformer.decoder.layers.1.norm3.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm1.weight": 256,
  "module.transformer.decoder.layers.2.norm1.bias": 256,
  "module.transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.catext_norm.weight": 256,
  "module.transformer.decoder.layers.2.catext_norm.bias": 256,
  "module.transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm2.weight": 256,
  "module.transformer.decoder.layers.2.norm2.bias": 256,
  "module.transformer.decoder.layers.2.linear1.weight": 524288,
  "module.transformer.decoder.layers.2.linear1.bias": 2048,
  "module.transformer.decoder.layers.2.linear2.weight": 524288,
  "module.transformer.decoder.layers.2.linear2.bias": 256,
  "module.transformer.decoder.layers.2.norm3.weight": 256,
  "module.transformer.decoder.layers.2.norm3.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm1.weight": 256,
  "module.transformer.decoder.layers.3.norm1.bias": 256,
  "module.transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.catext_norm.weight": 256,
  "module.transformer.decoder.layers.3.catext_norm.bias": 256,
  "module.transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm2.weight": 256,
  "module.transformer.decoder.layers.3.norm2.bias": 256,
  "module.transformer.decoder.layers.3.linear1.weight": 524288,
  "module.transformer.decoder.layers.3.linear1.bias": 2048,
  "module.transformer.decoder.layers.3.linear2.weight": 524288,
  "module.transformer.decoder.layers.3.linear2.bias": 256,
  "module.transformer.decoder.layers.3.norm3.weight": 256,
  "module.transformer.decoder.layers.3.norm3.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm1.weight": 256,
  "module.transformer.decoder.layers.4.norm1.bias": 256,
  "module.transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.catext_norm.weight": 256,
  "module.transformer.decoder.layers.4.catext_norm.bias": 256,
  "module.transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm2.weight": 256,
  "module.transformer.decoder.layers.4.norm2.bias": 256,
  "module.transformer.decoder.layers.4.linear1.weight": 524288,
  "module.transformer.decoder.layers.4.linear1.bias": 2048,
  "module.transformer.decoder.layers.4.linear2.weight": 524288,
  "module.transformer.decoder.layers.4.linear2.bias": 256,
  "module.transformer.decoder.layers.4.norm3.weight": 256,
  "module.transformer.decoder.layers.4.norm3.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm1.weight": 256,
  "module.transformer.decoder.layers.5.norm1.bias": 256,
  "module.transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.catext_norm.weight": 256,
  "module.transformer.decoder.layers.5.catext_norm.bias": 256,
  "module.transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm2.weight": 256,
  "module.transformer.decoder.layers.5.norm2.bias": 256,
  "module.transformer.decoder.layers.5.linear1.weight": 524288,
  "module.transformer.decoder.layers.5.linear1.bias": 2048,
  "module.transformer.decoder.layers.5.linear2.weight": 524288,
  "module.transformer.decoder.layers.5.linear2.bias": 256,
  "module.transformer.decoder.layers.5.norm3.weight": 256,
  "module.transformer.decoder.layers.5.norm3.bias": 256,
  "module.transformer.decoder.norm.weight": 256,
  "module.transformer.decoder.norm.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "module.transformer.decoder.ref_point_head.layers.0.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "module.transformer.decoder.ref_point_head.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "module.transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "module.transformer.tgt_embed.weight": 230400,
  "module.transformer.enc_output.weight": 65536,
  "module.transformer.enc_output.bias": 256,
  "module.transformer.enc_output_norm.weight": 256,
  "module.transformer.enc_output_norm.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "module.transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "module.feat_map.weight": 196608,
  "module.feat_map.bias": 256,
  "module.input_proj.0.0.weight": 49152,
  "module.input_proj.0.0.bias": 256,
  "module.input_proj.0.1.weight": 256,
  "module.input_proj.0.1.bias": 256,
  "module.input_proj.1.0.weight": 98304,
  "module.input_proj.1.0.bias": 256,
  "module.input_proj.1.1.weight": 256,
  "module.input_proj.1.1.bias": 256,
  "module.input_proj.2.0.weight": 196608,
  "module.input_proj.2.0.bias": 256,
  "module.input_proj.2.1.weight": 256,
  "module.input_proj.2.1.bias": 256,
  "module.input_proj.3.0.weight": 1769472,
  "module.input_proj.3.0.bias": 256,
  "module.input_proj.3.1.weight": 256,
  "module.input_proj.3.1.bias": 256,
  "module.backbone.0.patch_embed.proj.weight": 4608,
  "module.backbone.0.patch_embed.proj.bias": 96,
  "module.backbone.0.patch_embed.norm.weight": 96,
  "module.backbone.0.patch_embed.norm.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.downsample.reduction.weight": 73728,
  "module.backbone.0.layers.0.downsample.norm.weight": 384,
  "module.backbone.0.layers.0.downsample.norm.bias": 384,
  "module.backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.downsample.reduction.weight": 294912,
  "module.backbone.0.layers.1.downsample.norm.weight": 768,
  "module.backbone.0.layers.1.downsample.norm.bias": 768,
  "module.backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "module.backbone.0.layers.2.downsample.norm.weight": 1536,
  "module.backbone.0.layers.2.downsample.norm.bias": 1536,
  "module.backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "module.backbone.0.norm1.weight": 192,
  "module.backbone.0.norm1.bias": 192,
  "module.backbone.0.norm2.weight": 384,
  "module.backbone.0.norm2.bias": 384,
  "module.backbone.0.norm3.weight": 768,
  "module.backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-05 10:27:25,103 | [34mbuild dataset ... ...[0m
[32mINFO    [0m [32m2024-09-05 10:28:33,944 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-05 10:28:33,946 | [34mCommand: main.py --local-rank=0 --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --pretrain_model_path /home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-05 10:28:33,948 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-05 10:28:33,949 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-05 10:28:33,950 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:28:33,951 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:28:33,952 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='config/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_backend='nccl', dist_url='env://', distributed=True, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, gpu=0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path='/home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth', query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-05 10:28:33,954 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-05 10:28:46,017 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-05 10:28:46,152 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-05 10:28:46,164 | [34mparams before freezing:
{
  "module.transformer.level_embed": 1024,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.0.norm1.weight": 256,
  "module.transformer.encoder.layers.0.norm1.bias": 256,
  "module.transformer.encoder.layers.0.linear1.weight": 524288,
  "module.transformer.encoder.layers.0.linear1.bias": 2048,
  "module.transformer.encoder.layers.0.linear2.weight": 524288,
  "module.transformer.encoder.layers.0.linear2.bias": 256,
  "module.transformer.encoder.layers.0.norm2.weight": 256,
  "module.transformer.encoder.layers.0.norm2.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.1.norm1.weight": 256,
  "module.transformer.encoder.layers.1.norm1.bias": 256,
  "module.transformer.encoder.layers.1.linear1.weight": 524288,
  "module.transformer.encoder.layers.1.linear1.bias": 2048,
  "module.transformer.encoder.layers.1.linear2.weight": 524288,
  "module.transformer.encoder.layers.1.linear2.bias": 256,
  "module.transformer.encoder.layers.1.norm2.weight": 256,
  "module.transformer.encoder.layers.1.norm2.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.2.norm1.weight": 256,
  "module.transformer.encoder.layers.2.norm1.bias": 256,
  "module.transformer.encoder.layers.2.linear1.weight": 524288,
  "module.transformer.encoder.layers.2.linear1.bias": 2048,
  "module.transformer.encoder.layers.2.linear2.weight": 524288,
  "module.transformer.encoder.layers.2.linear2.bias": 256,
  "module.transformer.encoder.layers.2.norm2.weight": 256,
  "module.transformer.encoder.layers.2.norm2.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.3.norm1.weight": 256,
  "module.transformer.encoder.layers.3.norm1.bias": 256,
  "module.transformer.encoder.layers.3.linear1.weight": 524288,
  "module.transformer.encoder.layers.3.linear1.bias": 2048,
  "module.transformer.encoder.layers.3.linear2.weight": 524288,
  "module.transformer.encoder.layers.3.linear2.bias": 256,
  "module.transformer.encoder.layers.3.norm2.weight": 256,
  "module.transformer.encoder.layers.3.norm2.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.4.norm1.weight": 256,
  "module.transformer.encoder.layers.4.norm1.bias": 256,
  "module.transformer.encoder.layers.4.linear1.weight": 524288,
  "module.transformer.encoder.layers.4.linear1.bias": 2048,
  "module.transformer.encoder.layers.4.linear2.weight": 524288,
  "module.transformer.encoder.layers.4.linear2.bias": 256,
  "module.transformer.encoder.layers.4.norm2.weight": 256,
  "module.transformer.encoder.layers.4.norm2.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.5.norm1.weight": 256,
  "module.transformer.encoder.layers.5.norm1.bias": 256,
  "module.transformer.encoder.layers.5.linear1.weight": 524288,
  "module.transformer.encoder.layers.5.linear1.bias": 2048,
  "module.transformer.encoder.layers.5.linear2.weight": 524288,
  "module.transformer.encoder.layers.5.linear2.bias": 256,
  "module.transformer.encoder.layers.5.norm2.weight": 256,
  "module.transformer.encoder.layers.5.norm2.bias": 256,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.0.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.0.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear2.bias": 256,
  "module.transformer.encoder.text_layers.0.norm1.weight": 256,
  "module.transformer.encoder.text_layers.0.norm1.bias": 256,
  "module.transformer.encoder.text_layers.0.norm2.weight": 256,
  "module.transformer.encoder.text_layers.0.norm2.bias": 256,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.1.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.1.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear2.bias": 256,
  "module.transformer.encoder.text_layers.1.norm1.weight": 256,
  "module.transformer.encoder.text_layers.1.norm1.bias": 256,
  "module.transformer.encoder.text_layers.1.norm2.weight": 256,
  "module.transformer.encoder.text_layers.1.norm2.bias": 256,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.2.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.2.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear2.bias": 256,
  "module.transformer.encoder.text_layers.2.norm1.weight": 256,
  "module.transformer.encoder.text_layers.2.norm1.bias": 256,
  "module.transformer.encoder.text_layers.2.norm2.weight": 256,
  "module.transformer.encoder.text_layers.2.norm2.bias": 256,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.3.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.3.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear2.bias": 256,
  "module.transformer.encoder.text_layers.3.norm1.weight": 256,
  "module.transformer.encoder.text_layers.3.norm1.bias": 256,
  "module.transformer.encoder.text_layers.3.norm2.weight": 256,
  "module.transformer.encoder.text_layers.3.norm2.bias": 256,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.4.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.4.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear2.bias": 256,
  "module.transformer.encoder.text_layers.4.norm1.weight": 256,
  "module.transformer.encoder.text_layers.4.norm1.bias": 256,
  "module.transformer.encoder.text_layers.4.norm2.weight": 256,
  "module.transformer.encoder.text_layers.4.norm2.bias": 256,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.5.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.5.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear2.bias": 256,
  "module.transformer.encoder.text_layers.5.norm1.weight": 256,
  "module.transformer.encoder.text_layers.5.norm1.bias": 256,
  "module.transformer.encoder.text_layers.5.norm2.weight": 256,
  "module.transformer.encoder.text_layers.5.norm2.bias": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm1.weight": 256,
  "module.transformer.decoder.layers.0.norm1.bias": 256,
  "module.transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.catext_norm.weight": 256,
  "module.transformer.decoder.layers.0.catext_norm.bias": 256,
  "module.transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm2.weight": 256,
  "module.transformer.decoder.layers.0.norm2.bias": 256,
  "module.transformer.decoder.layers.0.linear1.weight": 524288,
  "module.transformer.decoder.layers.0.linear1.bias": 2048,
  "module.transformer.decoder.layers.0.linear2.weight": 524288,
  "module.transformer.decoder.layers.0.linear2.bias": 256,
  "module.transformer.decoder.layers.0.norm3.weight": 256,
  "module.transformer.decoder.layers.0.norm3.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm1.weight": 256,
  "module.transformer.decoder.layers.1.norm1.bias": 256,
  "module.transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.catext_norm.weight": 256,
  "module.transformer.decoder.layers.1.catext_norm.bias": 256,
  "module.transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm2.weight": 256,
  "module.transformer.decoder.layers.1.norm2.bias": 256,
  "module.transformer.decoder.layers.1.linear1.weight": 524288,
  "module.transformer.decoder.layers.1.linear1.bias": 2048,
  "module.transformer.decoder.layers.1.linear2.weight": 524288,
  "module.transformer.decoder.layers.1.linear2.bias": 256,
  "module.transformer.decoder.layers.1.norm3.weight": 256,
  "module.transformer.decoder.layers.1.norm3.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm1.weight": 256,
  "module.transformer.decoder.layers.2.norm1.bias": 256,
  "module.transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.catext_norm.weight": 256,
  "module.transformer.decoder.layers.2.catext_norm.bias": 256,
  "module.transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm2.weight": 256,
  "module.transformer.decoder.layers.2.norm2.bias": 256,
  "module.transformer.decoder.layers.2.linear1.weight": 524288,
  "module.transformer.decoder.layers.2.linear1.bias": 2048,
  "module.transformer.decoder.layers.2.linear2.weight": 524288,
  "module.transformer.decoder.layers.2.linear2.bias": 256,
  "module.transformer.decoder.layers.2.norm3.weight": 256,
  "module.transformer.decoder.layers.2.norm3.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm1.weight": 256,
  "module.transformer.decoder.layers.3.norm1.bias": 256,
  "module.transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.catext_norm.weight": 256,
  "module.transformer.decoder.layers.3.catext_norm.bias": 256,
  "module.transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm2.weight": 256,
  "module.transformer.decoder.layers.3.norm2.bias": 256,
  "module.transformer.decoder.layers.3.linear1.weight": 524288,
  "module.transformer.decoder.layers.3.linear1.bias": 2048,
  "module.transformer.decoder.layers.3.linear2.weight": 524288,
  "module.transformer.decoder.layers.3.linear2.bias": 256,
  "module.transformer.decoder.layers.3.norm3.weight": 256,
  "module.transformer.decoder.layers.3.norm3.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm1.weight": 256,
  "module.transformer.decoder.layers.4.norm1.bias": 256,
  "module.transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.catext_norm.weight": 256,
  "module.transformer.decoder.layers.4.catext_norm.bias": 256,
  "module.transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm2.weight": 256,
  "module.transformer.decoder.layers.4.norm2.bias": 256,
  "module.transformer.decoder.layers.4.linear1.weight": 524288,
  "module.transformer.decoder.layers.4.linear1.bias": 2048,
  "module.transformer.decoder.layers.4.linear2.weight": 524288,
  "module.transformer.decoder.layers.4.linear2.bias": 256,
  "module.transformer.decoder.layers.4.norm3.weight": 256,
  "module.transformer.decoder.layers.4.norm3.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm1.weight": 256,
  "module.transformer.decoder.layers.5.norm1.bias": 256,
  "module.transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.catext_norm.weight": 256,
  "module.transformer.decoder.layers.5.catext_norm.bias": 256,
  "module.transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm2.weight": 256,
  "module.transformer.decoder.layers.5.norm2.bias": 256,
  "module.transformer.decoder.layers.5.linear1.weight": 524288,
  "module.transformer.decoder.layers.5.linear1.bias": 2048,
  "module.transformer.decoder.layers.5.linear2.weight": 524288,
  "module.transformer.decoder.layers.5.linear2.bias": 256,
  "module.transformer.decoder.layers.5.norm3.weight": 256,
  "module.transformer.decoder.layers.5.norm3.bias": 256,
  "module.transformer.decoder.norm.weight": 256,
  "module.transformer.decoder.norm.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "module.transformer.decoder.ref_point_head.layers.0.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "module.transformer.decoder.ref_point_head.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "module.transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "module.transformer.tgt_embed.weight": 230400,
  "module.transformer.enc_output.weight": 65536,
  "module.transformer.enc_output.bias": 256,
  "module.transformer.enc_output_norm.weight": 256,
  "module.transformer.enc_output_norm.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "module.transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "module.bert.embeddings.word_embeddings.weight": 23440896,
  "module.bert.embeddings.position_embeddings.weight": 393216,
  "module.bert.embeddings.token_type_embeddings.weight": 1536,
  "module.bert.embeddings.LayerNorm.weight": 768,
  "module.bert.embeddings.LayerNorm.bias": 768,
  "module.bert.encoder.layer.0.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.query.bias": 768,
  "module.bert.encoder.layer.0.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.key.bias": 768,
  "module.bert.encoder.layer.0.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.value.bias": 768,
  "module.bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.0.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.0.output.dense.weight": 2359296,
  "module.bert.encoder.layer.0.output.dense.bias": 768,
  "module.bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.1.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.query.bias": 768,
  "module.bert.encoder.layer.1.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.key.bias": 768,
  "module.bert.encoder.layer.1.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.value.bias": 768,
  "module.bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.1.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.1.output.dense.weight": 2359296,
  "module.bert.encoder.layer.1.output.dense.bias": 768,
  "module.bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.2.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.query.bias": 768,
  "module.bert.encoder.layer.2.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.key.bias": 768,
  "module.bert.encoder.layer.2.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.value.bias": 768,
  "module.bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.2.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.2.output.dense.weight": 2359296,
  "module.bert.encoder.layer.2.output.dense.bias": 768,
  "module.bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.3.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.query.bias": 768,
  "module.bert.encoder.layer.3.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.key.bias": 768,
  "module.bert.encoder.layer.3.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.value.bias": 768,
  "module.bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.3.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.3.output.dense.weight": 2359296,
  "module.bert.encoder.layer.3.output.dense.bias": 768,
  "module.bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.4.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.query.bias": 768,
  "module.bert.encoder.layer.4.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.key.bias": 768,
  "module.bert.encoder.layer.4.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.value.bias": 768,
  "module.bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.4.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.4.output.dense.weight": 2359296,
  "module.bert.encoder.layer.4.output.dense.bias": 768,
  "module.bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.5.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.query.bias": 768,
  "module.bert.encoder.layer.5.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.key.bias": 768,
  "module.bert.encoder.layer.5.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.value.bias": 768,
  "module.bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.5.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.5.output.dense.weight": 2359296,
  "module.bert.encoder.layer.5.output.dense.bias": 768,
  "module.bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.6.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.query.bias": 768,
  "module.bert.encoder.layer.6.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.key.bias": 768,
  "module.bert.encoder.layer.6.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.value.bias": 768,
  "module.bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.6.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.6.output.dense.weight": 2359296,
  "module.bert.encoder.layer.6.output.dense.bias": 768,
  "module.bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.7.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.query.bias": 768,
  "module.bert.encoder.layer.7.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.key.bias": 768,
  "module.bert.encoder.layer.7.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.value.bias": 768,
  "module.bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.7.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.7.output.dense.weight": 2359296,
  "module.bert.encoder.layer.7.output.dense.bias": 768,
  "module.bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.8.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.query.bias": 768,
  "module.bert.encoder.layer.8.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.key.bias": 768,
  "module.bert.encoder.layer.8.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.value.bias": 768,
  "module.bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.8.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.8.output.dense.weight": 2359296,
  "module.bert.encoder.layer.8.output.dense.bias": 768,
  "module.bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.9.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.query.bias": 768,
  "module.bert.encoder.layer.9.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.key.bias": 768,
  "module.bert.encoder.layer.9.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.value.bias": 768,
  "module.bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.9.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.9.output.dense.weight": 2359296,
  "module.bert.encoder.layer.9.output.dense.bias": 768,
  "module.bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.10.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.query.bias": 768,
  "module.bert.encoder.layer.10.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.key.bias": 768,
  "module.bert.encoder.layer.10.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.value.bias": 768,
  "module.bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.10.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.10.output.dense.weight": 2359296,
  "module.bert.encoder.layer.10.output.dense.bias": 768,
  "module.bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.11.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.query.bias": 768,
  "module.bert.encoder.layer.11.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.key.bias": 768,
  "module.bert.encoder.layer.11.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.value.bias": 768,
  "module.bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.11.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.11.output.dense.weight": 2359296,
  "module.bert.encoder.layer.11.output.dense.bias": 768,
  "module.bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "module.feat_map.weight": 196608,
  "module.feat_map.bias": 256,
  "module.input_proj.0.0.weight": 49152,
  "module.input_proj.0.0.bias": 256,
  "module.input_proj.0.1.weight": 256,
  "module.input_proj.0.1.bias": 256,
  "module.input_proj.1.0.weight": 98304,
  "module.input_proj.1.0.bias": 256,
  "module.input_proj.1.1.weight": 256,
  "module.input_proj.1.1.bias": 256,
  "module.input_proj.2.0.weight": 196608,
  "module.input_proj.2.0.bias": 256,
  "module.input_proj.2.1.weight": 256,
  "module.input_proj.2.1.bias": 256,
  "module.input_proj.3.0.weight": 1769472,
  "module.input_proj.3.0.bias": 256,
  "module.input_proj.3.1.weight": 256,
  "module.input_proj.3.1.bias": 256,
  "module.backbone.0.patch_embed.proj.weight": 4608,
  "module.backbone.0.patch_embed.proj.bias": 96,
  "module.backbone.0.patch_embed.norm.weight": 96,
  "module.backbone.0.patch_embed.norm.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.downsample.reduction.weight": 73728,
  "module.backbone.0.layers.0.downsample.norm.weight": 384,
  "module.backbone.0.layers.0.downsample.norm.bias": 384,
  "module.backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.downsample.reduction.weight": 294912,
  "module.backbone.0.layers.1.downsample.norm.weight": 768,
  "module.backbone.0.layers.1.downsample.norm.bias": 768,
  "module.backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "module.backbone.0.layers.2.downsample.norm.weight": 1536,
  "module.backbone.0.layers.2.downsample.norm.bias": 1536,
  "module.backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "module.backbone.0.norm1.weight": 192,
  "module.backbone.0.norm1.bias": 192,
  "module.backbone.0.norm2.weight": 384,
  "module.backbone.0.norm2.bias": 384,
  "module.backbone.0.norm3.weight": 768,
  "module.backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-05 10:28:46,208 | [34mparams after freezing:
{
  "module.transformer.level_embed": 1024,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.0.norm1.weight": 256,
  "module.transformer.encoder.layers.0.norm1.bias": 256,
  "module.transformer.encoder.layers.0.linear1.weight": 524288,
  "module.transformer.encoder.layers.0.linear1.bias": 2048,
  "module.transformer.encoder.layers.0.linear2.weight": 524288,
  "module.transformer.encoder.layers.0.linear2.bias": 256,
  "module.transformer.encoder.layers.0.norm2.weight": 256,
  "module.transformer.encoder.layers.0.norm2.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.1.norm1.weight": 256,
  "module.transformer.encoder.layers.1.norm1.bias": 256,
  "module.transformer.encoder.layers.1.linear1.weight": 524288,
  "module.transformer.encoder.layers.1.linear1.bias": 2048,
  "module.transformer.encoder.layers.1.linear2.weight": 524288,
  "module.transformer.encoder.layers.1.linear2.bias": 256,
  "module.transformer.encoder.layers.1.norm2.weight": 256,
  "module.transformer.encoder.layers.1.norm2.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.2.norm1.weight": 256,
  "module.transformer.encoder.layers.2.norm1.bias": 256,
  "module.transformer.encoder.layers.2.linear1.weight": 524288,
  "module.transformer.encoder.layers.2.linear1.bias": 2048,
  "module.transformer.encoder.layers.2.linear2.weight": 524288,
  "module.transformer.encoder.layers.2.linear2.bias": 256,
  "module.transformer.encoder.layers.2.norm2.weight": 256,
  "module.transformer.encoder.layers.2.norm2.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.3.norm1.weight": 256,
  "module.transformer.encoder.layers.3.norm1.bias": 256,
  "module.transformer.encoder.layers.3.linear1.weight": 524288,
  "module.transformer.encoder.layers.3.linear1.bias": 2048,
  "module.transformer.encoder.layers.3.linear2.weight": 524288,
  "module.transformer.encoder.layers.3.linear2.bias": 256,
  "module.transformer.encoder.layers.3.norm2.weight": 256,
  "module.transformer.encoder.layers.3.norm2.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.4.norm1.weight": 256,
  "module.transformer.encoder.layers.4.norm1.bias": 256,
  "module.transformer.encoder.layers.4.linear1.weight": 524288,
  "module.transformer.encoder.layers.4.linear1.bias": 2048,
  "module.transformer.encoder.layers.4.linear2.weight": 524288,
  "module.transformer.encoder.layers.4.linear2.bias": 256,
  "module.transformer.encoder.layers.4.norm2.weight": 256,
  "module.transformer.encoder.layers.4.norm2.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.5.norm1.weight": 256,
  "module.transformer.encoder.layers.5.norm1.bias": 256,
  "module.transformer.encoder.layers.5.linear1.weight": 524288,
  "module.transformer.encoder.layers.5.linear1.bias": 2048,
  "module.transformer.encoder.layers.5.linear2.weight": 524288,
  "module.transformer.encoder.layers.5.linear2.bias": 256,
  "module.transformer.encoder.layers.5.norm2.weight": 256,
  "module.transformer.encoder.layers.5.norm2.bias": 256,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.0.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.0.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear2.bias": 256,
  "module.transformer.encoder.text_layers.0.norm1.weight": 256,
  "module.transformer.encoder.text_layers.0.norm1.bias": 256,
  "module.transformer.encoder.text_layers.0.norm2.weight": 256,
  "module.transformer.encoder.text_layers.0.norm2.bias": 256,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.1.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.1.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear2.bias": 256,
  "module.transformer.encoder.text_layers.1.norm1.weight": 256,
  "module.transformer.encoder.text_layers.1.norm1.bias": 256,
  "module.transformer.encoder.text_layers.1.norm2.weight": 256,
  "module.transformer.encoder.text_layers.1.norm2.bias": 256,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.2.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.2.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear2.bias": 256,
  "module.transformer.encoder.text_layers.2.norm1.weight": 256,
  "module.transformer.encoder.text_layers.2.norm1.bias": 256,
  "module.transformer.encoder.text_layers.2.norm2.weight": 256,
  "module.transformer.encoder.text_layers.2.norm2.bias": 256,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.3.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.3.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear2.bias": 256,
  "module.transformer.encoder.text_layers.3.norm1.weight": 256,
  "module.transformer.encoder.text_layers.3.norm1.bias": 256,
  "module.transformer.encoder.text_layers.3.norm2.weight": 256,
  "module.transformer.encoder.text_layers.3.norm2.bias": 256,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.4.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.4.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear2.bias": 256,
  "module.transformer.encoder.text_layers.4.norm1.weight": 256,
  "module.transformer.encoder.text_layers.4.norm1.bias": 256,
  "module.transformer.encoder.text_layers.4.norm2.weight": 256,
  "module.transformer.encoder.text_layers.4.norm2.bias": 256,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.5.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.5.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear2.bias": 256,
  "module.transformer.encoder.text_layers.5.norm1.weight": 256,
  "module.transformer.encoder.text_layers.5.norm1.bias": 256,
  "module.transformer.encoder.text_layers.5.norm2.weight": 256,
  "module.transformer.encoder.text_layers.5.norm2.bias": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm1.weight": 256,
  "module.transformer.decoder.layers.0.norm1.bias": 256,
  "module.transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.catext_norm.weight": 256,
  "module.transformer.decoder.layers.0.catext_norm.bias": 256,
  "module.transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm2.weight": 256,
  "module.transformer.decoder.layers.0.norm2.bias": 256,
  "module.transformer.decoder.layers.0.linear1.weight": 524288,
  "module.transformer.decoder.layers.0.linear1.bias": 2048,
  "module.transformer.decoder.layers.0.linear2.weight": 524288,
  "module.transformer.decoder.layers.0.linear2.bias": 256,
  "module.transformer.decoder.layers.0.norm3.weight": 256,
  "module.transformer.decoder.layers.0.norm3.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm1.weight": 256,
  "module.transformer.decoder.layers.1.norm1.bias": 256,
  "module.transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.catext_norm.weight": 256,
  "module.transformer.decoder.layers.1.catext_norm.bias": 256,
  "module.transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm2.weight": 256,
  "module.transformer.decoder.layers.1.norm2.bias": 256,
  "module.transformer.decoder.layers.1.linear1.weight": 524288,
  "module.transformer.decoder.layers.1.linear1.bias": 2048,
  "module.transformer.decoder.layers.1.linear2.weight": 524288,
  "module.transformer.decoder.layers.1.linear2.bias": 256,
  "module.transformer.decoder.layers.1.norm3.weight": 256,
  "module.transformer.decoder.layers.1.norm3.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm1.weight": 256,
  "module.transformer.decoder.layers.2.norm1.bias": 256,
  "module.transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.catext_norm.weight": 256,
  "module.transformer.decoder.layers.2.catext_norm.bias": 256,
  "module.transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm2.weight": 256,
  "module.transformer.decoder.layers.2.norm2.bias": 256,
  "module.transformer.decoder.layers.2.linear1.weight": 524288,
  "module.transformer.decoder.layers.2.linear1.bias": 2048,
  "module.transformer.decoder.layers.2.linear2.weight": 524288,
  "module.transformer.decoder.layers.2.linear2.bias": 256,
  "module.transformer.decoder.layers.2.norm3.weight": 256,
  "module.transformer.decoder.layers.2.norm3.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm1.weight": 256,
  "module.transformer.decoder.layers.3.norm1.bias": 256,
  "module.transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.catext_norm.weight": 256,
  "module.transformer.decoder.layers.3.catext_norm.bias": 256,
  "module.transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm2.weight": 256,
  "module.transformer.decoder.layers.3.norm2.bias": 256,
  "module.transformer.decoder.layers.3.linear1.weight": 524288,
  "module.transformer.decoder.layers.3.linear1.bias": 2048,
  "module.transformer.decoder.layers.3.linear2.weight": 524288,
  "module.transformer.decoder.layers.3.linear2.bias": 256,
  "module.transformer.decoder.layers.3.norm3.weight": 256,
  "module.transformer.decoder.layers.3.norm3.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm1.weight": 256,
  "module.transformer.decoder.layers.4.norm1.bias": 256,
  "module.transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.catext_norm.weight": 256,
  "module.transformer.decoder.layers.4.catext_norm.bias": 256,
  "module.transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm2.weight": 256,
  "module.transformer.decoder.layers.4.norm2.bias": 256,
  "module.transformer.decoder.layers.4.linear1.weight": 524288,
  "module.transformer.decoder.layers.4.linear1.bias": 2048,
  "module.transformer.decoder.layers.4.linear2.weight": 524288,
  "module.transformer.decoder.layers.4.linear2.bias": 256,
  "module.transformer.decoder.layers.4.norm3.weight": 256,
  "module.transformer.decoder.layers.4.norm3.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm1.weight": 256,
  "module.transformer.decoder.layers.5.norm1.bias": 256,
  "module.transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.catext_norm.weight": 256,
  "module.transformer.decoder.layers.5.catext_norm.bias": 256,
  "module.transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm2.weight": 256,
  "module.transformer.decoder.layers.5.norm2.bias": 256,
  "module.transformer.decoder.layers.5.linear1.weight": 524288,
  "module.transformer.decoder.layers.5.linear1.bias": 2048,
  "module.transformer.decoder.layers.5.linear2.weight": 524288,
  "module.transformer.decoder.layers.5.linear2.bias": 256,
  "module.transformer.decoder.layers.5.norm3.weight": 256,
  "module.transformer.decoder.layers.5.norm3.bias": 256,
  "module.transformer.decoder.norm.weight": 256,
  "module.transformer.decoder.norm.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "module.transformer.decoder.ref_point_head.layers.0.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "module.transformer.decoder.ref_point_head.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "module.transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "module.transformer.tgt_embed.weight": 230400,
  "module.transformer.enc_output.weight": 65536,
  "module.transformer.enc_output.bias": 256,
  "module.transformer.enc_output_norm.weight": 256,
  "module.transformer.enc_output_norm.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "module.transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "module.feat_map.weight": 196608,
  "module.feat_map.bias": 256,
  "module.input_proj.0.0.weight": 49152,
  "module.input_proj.0.0.bias": 256,
  "module.input_proj.0.1.weight": 256,
  "module.input_proj.0.1.bias": 256,
  "module.input_proj.1.0.weight": 98304,
  "module.input_proj.1.0.bias": 256,
  "module.input_proj.1.1.weight": 256,
  "module.input_proj.1.1.bias": 256,
  "module.input_proj.2.0.weight": 196608,
  "module.input_proj.2.0.bias": 256,
  "module.input_proj.2.1.weight": 256,
  "module.input_proj.2.1.bias": 256,
  "module.input_proj.3.0.weight": 1769472,
  "module.input_proj.3.0.bias": 256,
  "module.input_proj.3.1.weight": 256,
  "module.input_proj.3.1.bias": 256,
  "module.backbone.0.patch_embed.proj.weight": 4608,
  "module.backbone.0.patch_embed.proj.bias": 96,
  "module.backbone.0.patch_embed.norm.weight": 96,
  "module.backbone.0.patch_embed.norm.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.downsample.reduction.weight": 73728,
  "module.backbone.0.layers.0.downsample.norm.weight": 384,
  "module.backbone.0.layers.0.downsample.norm.bias": 384,
  "module.backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.downsample.reduction.weight": 294912,
  "module.backbone.0.layers.1.downsample.norm.weight": 768,
  "module.backbone.0.layers.1.downsample.norm.bias": 768,
  "module.backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "module.backbone.0.layers.2.downsample.norm.weight": 1536,
  "module.backbone.0.layers.2.downsample.norm.bias": 1536,
  "module.backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "module.backbone.0.norm1.weight": 192,
  "module.backbone.0.norm1.bias": 192,
  "module.backbone.0.norm2.weight": 384,
  "module.backbone.0.norm2.bias": 384,
  "module.backbone.0.norm3.weight": 768,
  "module.backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-05 10:28:46,220 | [34mbuild dataset ... ...[0m
[32mINFO    [0m [32m2024-09-05 10:30:31,128 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-05 10:30:31,129 | [34mCommand: main.py --local-rank=0 --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --pretrain_model_path /home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-05 10:30:31,131 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-05 10:30:31,132 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-05 10:30:31,132 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:30:31,133 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-05 10:30:31,133 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='config/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_backend='nccl', dist_url='env://', distributed=True, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, gpu=0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path='/home/jiask/Open-GroundingDino-main/weights/groundingdino_swint_ogc.pth', query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-05 10:30:31,134 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-05 10:30:41,782 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-05 10:30:41,898 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-05 10:30:41,907 | [34mparams before freezing:
{
  "module.transformer.level_embed": 1024,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.0.norm1.weight": 256,
  "module.transformer.encoder.layers.0.norm1.bias": 256,
  "module.transformer.encoder.layers.0.linear1.weight": 524288,
  "module.transformer.encoder.layers.0.linear1.bias": 2048,
  "module.transformer.encoder.layers.0.linear2.weight": 524288,
  "module.transformer.encoder.layers.0.linear2.bias": 256,
  "module.transformer.encoder.layers.0.norm2.weight": 256,
  "module.transformer.encoder.layers.0.norm2.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.1.norm1.weight": 256,
  "module.transformer.encoder.layers.1.norm1.bias": 256,
  "module.transformer.encoder.layers.1.linear1.weight": 524288,
  "module.transformer.encoder.layers.1.linear1.bias": 2048,
  "module.transformer.encoder.layers.1.linear2.weight": 524288,
  "module.transformer.encoder.layers.1.linear2.bias": 256,
  "module.transformer.encoder.layers.1.norm2.weight": 256,
  "module.transformer.encoder.layers.1.norm2.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.2.norm1.weight": 256,
  "module.transformer.encoder.layers.2.norm1.bias": 256,
  "module.transformer.encoder.layers.2.linear1.weight": 524288,
  "module.transformer.encoder.layers.2.linear1.bias": 2048,
  "module.transformer.encoder.layers.2.linear2.weight": 524288,
  "module.transformer.encoder.layers.2.linear2.bias": 256,
  "module.transformer.encoder.layers.2.norm2.weight": 256,
  "module.transformer.encoder.layers.2.norm2.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.3.norm1.weight": 256,
  "module.transformer.encoder.layers.3.norm1.bias": 256,
  "module.transformer.encoder.layers.3.linear1.weight": 524288,
  "module.transformer.encoder.layers.3.linear1.bias": 2048,
  "module.transformer.encoder.layers.3.linear2.weight": 524288,
  "module.transformer.encoder.layers.3.linear2.bias": 256,
  "module.transformer.encoder.layers.3.norm2.weight": 256,
  "module.transformer.encoder.layers.3.norm2.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.4.norm1.weight": 256,
  "module.transformer.encoder.layers.4.norm1.bias": 256,
  "module.transformer.encoder.layers.4.linear1.weight": 524288,
  "module.transformer.encoder.layers.4.linear1.bias": 2048,
  "module.transformer.encoder.layers.4.linear2.weight": 524288,
  "module.transformer.encoder.layers.4.linear2.bias": 256,
  "module.transformer.encoder.layers.4.norm2.weight": 256,
  "module.transformer.encoder.layers.4.norm2.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.5.norm1.weight": 256,
  "module.transformer.encoder.layers.5.norm1.bias": 256,
  "module.transformer.encoder.layers.5.linear1.weight": 524288,
  "module.transformer.encoder.layers.5.linear1.bias": 2048,
  "module.transformer.encoder.layers.5.linear2.weight": 524288,
  "module.transformer.encoder.layers.5.linear2.bias": 256,
  "module.transformer.encoder.layers.5.norm2.weight": 256,
  "module.transformer.encoder.layers.5.norm2.bias": 256,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.0.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.0.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear2.bias": 256,
  "module.transformer.encoder.text_layers.0.norm1.weight": 256,
  "module.transformer.encoder.text_layers.0.norm1.bias": 256,
  "module.transformer.encoder.text_layers.0.norm2.weight": 256,
  "module.transformer.encoder.text_layers.0.norm2.bias": 256,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.1.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.1.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear2.bias": 256,
  "module.transformer.encoder.text_layers.1.norm1.weight": 256,
  "module.transformer.encoder.text_layers.1.norm1.bias": 256,
  "module.transformer.encoder.text_layers.1.norm2.weight": 256,
  "module.transformer.encoder.text_layers.1.norm2.bias": 256,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.2.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.2.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear2.bias": 256,
  "module.transformer.encoder.text_layers.2.norm1.weight": 256,
  "module.transformer.encoder.text_layers.2.norm1.bias": 256,
  "module.transformer.encoder.text_layers.2.norm2.weight": 256,
  "module.transformer.encoder.text_layers.2.norm2.bias": 256,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.3.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.3.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear2.bias": 256,
  "module.transformer.encoder.text_layers.3.norm1.weight": 256,
  "module.transformer.encoder.text_layers.3.norm1.bias": 256,
  "module.transformer.encoder.text_layers.3.norm2.weight": 256,
  "module.transformer.encoder.text_layers.3.norm2.bias": 256,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.4.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.4.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear2.bias": 256,
  "module.transformer.encoder.text_layers.4.norm1.weight": 256,
  "module.transformer.encoder.text_layers.4.norm1.bias": 256,
  "module.transformer.encoder.text_layers.4.norm2.weight": 256,
  "module.transformer.encoder.text_layers.4.norm2.bias": 256,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.5.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.5.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear2.bias": 256,
  "module.transformer.encoder.text_layers.5.norm1.weight": 256,
  "module.transformer.encoder.text_layers.5.norm1.bias": 256,
  "module.transformer.encoder.text_layers.5.norm2.weight": 256,
  "module.transformer.encoder.text_layers.5.norm2.bias": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm1.weight": 256,
  "module.transformer.decoder.layers.0.norm1.bias": 256,
  "module.transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.catext_norm.weight": 256,
  "module.transformer.decoder.layers.0.catext_norm.bias": 256,
  "module.transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm2.weight": 256,
  "module.transformer.decoder.layers.0.norm2.bias": 256,
  "module.transformer.decoder.layers.0.linear1.weight": 524288,
  "module.transformer.decoder.layers.0.linear1.bias": 2048,
  "module.transformer.decoder.layers.0.linear2.weight": 524288,
  "module.transformer.decoder.layers.0.linear2.bias": 256,
  "module.transformer.decoder.layers.0.norm3.weight": 256,
  "module.transformer.decoder.layers.0.norm3.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm1.weight": 256,
  "module.transformer.decoder.layers.1.norm1.bias": 256,
  "module.transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.catext_norm.weight": 256,
  "module.transformer.decoder.layers.1.catext_norm.bias": 256,
  "module.transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm2.weight": 256,
  "module.transformer.decoder.layers.1.norm2.bias": 256,
  "module.transformer.decoder.layers.1.linear1.weight": 524288,
  "module.transformer.decoder.layers.1.linear1.bias": 2048,
  "module.transformer.decoder.layers.1.linear2.weight": 524288,
  "module.transformer.decoder.layers.1.linear2.bias": 256,
  "module.transformer.decoder.layers.1.norm3.weight": 256,
  "module.transformer.decoder.layers.1.norm3.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm1.weight": 256,
  "module.transformer.decoder.layers.2.norm1.bias": 256,
  "module.transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.catext_norm.weight": 256,
  "module.transformer.decoder.layers.2.catext_norm.bias": 256,
  "module.transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm2.weight": 256,
  "module.transformer.decoder.layers.2.norm2.bias": 256,
  "module.transformer.decoder.layers.2.linear1.weight": 524288,
  "module.transformer.decoder.layers.2.linear1.bias": 2048,
  "module.transformer.decoder.layers.2.linear2.weight": 524288,
  "module.transformer.decoder.layers.2.linear2.bias": 256,
  "module.transformer.decoder.layers.2.norm3.weight": 256,
  "module.transformer.decoder.layers.2.norm3.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm1.weight": 256,
  "module.transformer.decoder.layers.3.norm1.bias": 256,
  "module.transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.catext_norm.weight": 256,
  "module.transformer.decoder.layers.3.catext_norm.bias": 256,
  "module.transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm2.weight": 256,
  "module.transformer.decoder.layers.3.norm2.bias": 256,
  "module.transformer.decoder.layers.3.linear1.weight": 524288,
  "module.transformer.decoder.layers.3.linear1.bias": 2048,
  "module.transformer.decoder.layers.3.linear2.weight": 524288,
  "module.transformer.decoder.layers.3.linear2.bias": 256,
  "module.transformer.decoder.layers.3.norm3.weight": 256,
  "module.transformer.decoder.layers.3.norm3.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm1.weight": 256,
  "module.transformer.decoder.layers.4.norm1.bias": 256,
  "module.transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.catext_norm.weight": 256,
  "module.transformer.decoder.layers.4.catext_norm.bias": 256,
  "module.transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm2.weight": 256,
  "module.transformer.decoder.layers.4.norm2.bias": 256,
  "module.transformer.decoder.layers.4.linear1.weight": 524288,
  "module.transformer.decoder.layers.4.linear1.bias": 2048,
  "module.transformer.decoder.layers.4.linear2.weight": 524288,
  "module.transformer.decoder.layers.4.linear2.bias": 256,
  "module.transformer.decoder.layers.4.norm3.weight": 256,
  "module.transformer.decoder.layers.4.norm3.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm1.weight": 256,
  "module.transformer.decoder.layers.5.norm1.bias": 256,
  "module.transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.catext_norm.weight": 256,
  "module.transformer.decoder.layers.5.catext_norm.bias": 256,
  "module.transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm2.weight": 256,
  "module.transformer.decoder.layers.5.norm2.bias": 256,
  "module.transformer.decoder.layers.5.linear1.weight": 524288,
  "module.transformer.decoder.layers.5.linear1.bias": 2048,
  "module.transformer.decoder.layers.5.linear2.weight": 524288,
  "module.transformer.decoder.layers.5.linear2.bias": 256,
  "module.transformer.decoder.layers.5.norm3.weight": 256,
  "module.transformer.decoder.layers.5.norm3.bias": 256,
  "module.transformer.decoder.norm.weight": 256,
  "module.transformer.decoder.norm.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "module.transformer.decoder.ref_point_head.layers.0.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "module.transformer.decoder.ref_point_head.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "module.transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "module.transformer.tgt_embed.weight": 230400,
  "module.transformer.enc_output.weight": 65536,
  "module.transformer.enc_output.bias": 256,
  "module.transformer.enc_output_norm.weight": 256,
  "module.transformer.enc_output_norm.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "module.transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "module.bert.embeddings.word_embeddings.weight": 23440896,
  "module.bert.embeddings.position_embeddings.weight": 393216,
  "module.bert.embeddings.token_type_embeddings.weight": 1536,
  "module.bert.embeddings.LayerNorm.weight": 768,
  "module.bert.embeddings.LayerNorm.bias": 768,
  "module.bert.encoder.layer.0.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.query.bias": 768,
  "module.bert.encoder.layer.0.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.key.bias": 768,
  "module.bert.encoder.layer.0.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.0.attention.self.value.bias": 768,
  "module.bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.0.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.0.output.dense.weight": 2359296,
  "module.bert.encoder.layer.0.output.dense.bias": 768,
  "module.bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.1.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.query.bias": 768,
  "module.bert.encoder.layer.1.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.key.bias": 768,
  "module.bert.encoder.layer.1.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.1.attention.self.value.bias": 768,
  "module.bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.1.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.1.output.dense.weight": 2359296,
  "module.bert.encoder.layer.1.output.dense.bias": 768,
  "module.bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.2.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.query.bias": 768,
  "module.bert.encoder.layer.2.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.key.bias": 768,
  "module.bert.encoder.layer.2.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.2.attention.self.value.bias": 768,
  "module.bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.2.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.2.output.dense.weight": 2359296,
  "module.bert.encoder.layer.2.output.dense.bias": 768,
  "module.bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.3.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.query.bias": 768,
  "module.bert.encoder.layer.3.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.key.bias": 768,
  "module.bert.encoder.layer.3.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.3.attention.self.value.bias": 768,
  "module.bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.3.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.3.output.dense.weight": 2359296,
  "module.bert.encoder.layer.3.output.dense.bias": 768,
  "module.bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.4.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.query.bias": 768,
  "module.bert.encoder.layer.4.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.key.bias": 768,
  "module.bert.encoder.layer.4.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.4.attention.self.value.bias": 768,
  "module.bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.4.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.4.output.dense.weight": 2359296,
  "module.bert.encoder.layer.4.output.dense.bias": 768,
  "module.bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.5.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.query.bias": 768,
  "module.bert.encoder.layer.5.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.key.bias": 768,
  "module.bert.encoder.layer.5.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.5.attention.self.value.bias": 768,
  "module.bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.5.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.5.output.dense.weight": 2359296,
  "module.bert.encoder.layer.5.output.dense.bias": 768,
  "module.bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.6.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.query.bias": 768,
  "module.bert.encoder.layer.6.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.key.bias": 768,
  "module.bert.encoder.layer.6.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.6.attention.self.value.bias": 768,
  "module.bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.6.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.6.output.dense.weight": 2359296,
  "module.bert.encoder.layer.6.output.dense.bias": 768,
  "module.bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.7.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.query.bias": 768,
  "module.bert.encoder.layer.7.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.key.bias": 768,
  "module.bert.encoder.layer.7.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.7.attention.self.value.bias": 768,
  "module.bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.7.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.7.output.dense.weight": 2359296,
  "module.bert.encoder.layer.7.output.dense.bias": 768,
  "module.bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.8.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.query.bias": 768,
  "module.bert.encoder.layer.8.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.key.bias": 768,
  "module.bert.encoder.layer.8.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.8.attention.self.value.bias": 768,
  "module.bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.8.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.8.output.dense.weight": 2359296,
  "module.bert.encoder.layer.8.output.dense.bias": 768,
  "module.bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.9.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.query.bias": 768,
  "module.bert.encoder.layer.9.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.key.bias": 768,
  "module.bert.encoder.layer.9.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.9.attention.self.value.bias": 768,
  "module.bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.9.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.9.output.dense.weight": 2359296,
  "module.bert.encoder.layer.9.output.dense.bias": 768,
  "module.bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.10.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.query.bias": 768,
  "module.bert.encoder.layer.10.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.key.bias": 768,
  "module.bert.encoder.layer.10.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.10.attention.self.value.bias": 768,
  "module.bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.10.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.10.output.dense.weight": 2359296,
  "module.bert.encoder.layer.10.output.dense.bias": 768,
  "module.bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.11.attention.self.query.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.query.bias": 768,
  "module.bert.encoder.layer.11.attention.self.key.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.key.bias": 768,
  "module.bert.encoder.layer.11.attention.self.value.weight": 589824,
  "module.bert.encoder.layer.11.attention.self.value.bias": 768,
  "module.bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "module.bert.encoder.layer.11.attention.output.dense.bias": 768,
  "module.bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "module.bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "module.bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "module.bert.encoder.layer.11.output.dense.weight": 2359296,
  "module.bert.encoder.layer.11.output.dense.bias": 768,
  "module.bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "module.bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "module.feat_map.weight": 196608,
  "module.feat_map.bias": 256,
  "module.input_proj.0.0.weight": 49152,
  "module.input_proj.0.0.bias": 256,
  "module.input_proj.0.1.weight": 256,
  "module.input_proj.0.1.bias": 256,
  "module.input_proj.1.0.weight": 98304,
  "module.input_proj.1.0.bias": 256,
  "module.input_proj.1.1.weight": 256,
  "module.input_proj.1.1.bias": 256,
  "module.input_proj.2.0.weight": 196608,
  "module.input_proj.2.0.bias": 256,
  "module.input_proj.2.1.weight": 256,
  "module.input_proj.2.1.bias": 256,
  "module.input_proj.3.0.weight": 1769472,
  "module.input_proj.3.0.bias": 256,
  "module.input_proj.3.1.weight": 256,
  "module.input_proj.3.1.bias": 256,
  "module.backbone.0.patch_embed.proj.weight": 4608,
  "module.backbone.0.patch_embed.proj.bias": 96,
  "module.backbone.0.patch_embed.norm.weight": 96,
  "module.backbone.0.patch_embed.norm.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.downsample.reduction.weight": 73728,
  "module.backbone.0.layers.0.downsample.norm.weight": 384,
  "module.backbone.0.layers.0.downsample.norm.bias": 384,
  "module.backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.downsample.reduction.weight": 294912,
  "module.backbone.0.layers.1.downsample.norm.weight": 768,
  "module.backbone.0.layers.1.downsample.norm.bias": 768,
  "module.backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "module.backbone.0.layers.2.downsample.norm.weight": 1536,
  "module.backbone.0.layers.2.downsample.norm.bias": 1536,
  "module.backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "module.backbone.0.norm1.weight": 192,
  "module.backbone.0.norm1.bias": 192,
  "module.backbone.0.norm2.weight": 384,
  "module.backbone.0.norm2.bias": 384,
  "module.backbone.0.norm3.weight": 768,
  "module.backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-05 10:30:41,946 | [34mparams after freezing:
{
  "module.transformer.level_embed": 1024,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.0.norm1.weight": 256,
  "module.transformer.encoder.layers.0.norm1.bias": 256,
  "module.transformer.encoder.layers.0.linear1.weight": 524288,
  "module.transformer.encoder.layers.0.linear1.bias": 2048,
  "module.transformer.encoder.layers.0.linear2.weight": 524288,
  "module.transformer.encoder.layers.0.linear2.bias": 256,
  "module.transformer.encoder.layers.0.norm2.weight": 256,
  "module.transformer.encoder.layers.0.norm2.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.1.norm1.weight": 256,
  "module.transformer.encoder.layers.1.norm1.bias": 256,
  "module.transformer.encoder.layers.1.linear1.weight": 524288,
  "module.transformer.encoder.layers.1.linear1.bias": 2048,
  "module.transformer.encoder.layers.1.linear2.weight": 524288,
  "module.transformer.encoder.layers.1.linear2.bias": 256,
  "module.transformer.encoder.layers.1.norm2.weight": 256,
  "module.transformer.encoder.layers.1.norm2.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.2.norm1.weight": 256,
  "module.transformer.encoder.layers.2.norm1.bias": 256,
  "module.transformer.encoder.layers.2.linear1.weight": 524288,
  "module.transformer.encoder.layers.2.linear1.bias": 2048,
  "module.transformer.encoder.layers.2.linear2.weight": 524288,
  "module.transformer.encoder.layers.2.linear2.bias": 256,
  "module.transformer.encoder.layers.2.norm2.weight": 256,
  "module.transformer.encoder.layers.2.norm2.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.3.norm1.weight": 256,
  "module.transformer.encoder.layers.3.norm1.bias": 256,
  "module.transformer.encoder.layers.3.linear1.weight": 524288,
  "module.transformer.encoder.layers.3.linear1.bias": 2048,
  "module.transformer.encoder.layers.3.linear2.weight": 524288,
  "module.transformer.encoder.layers.3.linear2.bias": 256,
  "module.transformer.encoder.layers.3.norm2.weight": 256,
  "module.transformer.encoder.layers.3.norm2.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.4.norm1.weight": 256,
  "module.transformer.encoder.layers.4.norm1.bias": 256,
  "module.transformer.encoder.layers.4.linear1.weight": 524288,
  "module.transformer.encoder.layers.4.linear1.bias": 2048,
  "module.transformer.encoder.layers.4.linear2.weight": 524288,
  "module.transformer.encoder.layers.4.linear2.bias": 256,
  "module.transformer.encoder.layers.4.norm2.weight": 256,
  "module.transformer.encoder.layers.4.norm2.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "module.transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "module.transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "module.transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "module.transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "module.transformer.encoder.layers.5.norm1.weight": 256,
  "module.transformer.encoder.layers.5.norm1.bias": 256,
  "module.transformer.encoder.layers.5.linear1.weight": 524288,
  "module.transformer.encoder.layers.5.linear1.bias": 2048,
  "module.transformer.encoder.layers.5.linear2.weight": 524288,
  "module.transformer.encoder.layers.5.linear2.bias": 256,
  "module.transformer.encoder.layers.5.norm2.weight": 256,
  "module.transformer.encoder.layers.5.norm2.bias": 256,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.0.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.0.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.0.linear2.bias": 256,
  "module.transformer.encoder.text_layers.0.norm1.weight": 256,
  "module.transformer.encoder.text_layers.0.norm1.bias": 256,
  "module.transformer.encoder.text_layers.0.norm2.weight": 256,
  "module.transformer.encoder.text_layers.0.norm2.bias": 256,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.1.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.1.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.1.linear2.bias": 256,
  "module.transformer.encoder.text_layers.1.norm1.weight": 256,
  "module.transformer.encoder.text_layers.1.norm1.bias": 256,
  "module.transformer.encoder.text_layers.1.norm2.weight": 256,
  "module.transformer.encoder.text_layers.1.norm2.bias": 256,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.2.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.2.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.2.linear2.bias": 256,
  "module.transformer.encoder.text_layers.2.norm1.weight": 256,
  "module.transformer.encoder.text_layers.2.norm1.bias": 256,
  "module.transformer.encoder.text_layers.2.norm2.weight": 256,
  "module.transformer.encoder.text_layers.2.norm2.bias": 256,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.3.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.3.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.3.linear2.bias": 256,
  "module.transformer.encoder.text_layers.3.norm1.weight": 256,
  "module.transformer.encoder.text_layers.3.norm1.bias": 256,
  "module.transformer.encoder.text_layers.3.norm2.weight": 256,
  "module.transformer.encoder.text_layers.3.norm2.bias": 256,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.4.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.4.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.4.linear2.bias": 256,
  "module.transformer.encoder.text_layers.4.norm1.weight": 256,
  "module.transformer.encoder.text_layers.4.norm1.bias": 256,
  "module.transformer.encoder.text_layers.4.norm2.weight": 256,
  "module.transformer.encoder.text_layers.4.norm2.bias": 256,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.encoder.text_layers.5.linear1.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear1.bias": 1024,
  "module.transformer.encoder.text_layers.5.linear2.weight": 262144,
  "module.transformer.encoder.text_layers.5.linear2.bias": 256,
  "module.transformer.encoder.text_layers.5.norm1.weight": 256,
  "module.transformer.encoder.text_layers.5.norm1.bias": 256,
  "module.transformer.encoder.text_layers.5.norm2.weight": 256,
  "module.transformer.encoder.text_layers.5.norm2.bias": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.0.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.1.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.2.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.3.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.4.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_v": 256,
  "module.transformer.encoder.fusion_layers.5.gamma_l": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "module.transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "module.transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm1.weight": 256,
  "module.transformer.decoder.layers.0.norm1.bias": 256,
  "module.transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.catext_norm.weight": 256,
  "module.transformer.decoder.layers.0.catext_norm.bias": 256,
  "module.transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.0.norm2.weight": 256,
  "module.transformer.decoder.layers.0.norm2.bias": 256,
  "module.transformer.decoder.layers.0.linear1.weight": 524288,
  "module.transformer.decoder.layers.0.linear1.bias": 2048,
  "module.transformer.decoder.layers.0.linear2.weight": 524288,
  "module.transformer.decoder.layers.0.linear2.bias": 256,
  "module.transformer.decoder.layers.0.norm3.weight": 256,
  "module.transformer.decoder.layers.0.norm3.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm1.weight": 256,
  "module.transformer.decoder.layers.1.norm1.bias": 256,
  "module.transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.catext_norm.weight": 256,
  "module.transformer.decoder.layers.1.catext_norm.bias": 256,
  "module.transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.1.norm2.weight": 256,
  "module.transformer.decoder.layers.1.norm2.bias": 256,
  "module.transformer.decoder.layers.1.linear1.weight": 524288,
  "module.transformer.decoder.layers.1.linear1.bias": 2048,
  "module.transformer.decoder.layers.1.linear2.weight": 524288,
  "module.transformer.decoder.layers.1.linear2.bias": 256,
  "module.transformer.decoder.layers.1.norm3.weight": 256,
  "module.transformer.decoder.layers.1.norm3.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm1.weight": 256,
  "module.transformer.decoder.layers.2.norm1.bias": 256,
  "module.transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.catext_norm.weight": 256,
  "module.transformer.decoder.layers.2.catext_norm.bias": 256,
  "module.transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.2.norm2.weight": 256,
  "module.transformer.decoder.layers.2.norm2.bias": 256,
  "module.transformer.decoder.layers.2.linear1.weight": 524288,
  "module.transformer.decoder.layers.2.linear1.bias": 2048,
  "module.transformer.decoder.layers.2.linear2.weight": 524288,
  "module.transformer.decoder.layers.2.linear2.bias": 256,
  "module.transformer.decoder.layers.2.norm3.weight": 256,
  "module.transformer.decoder.layers.2.norm3.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm1.weight": 256,
  "module.transformer.decoder.layers.3.norm1.bias": 256,
  "module.transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.catext_norm.weight": 256,
  "module.transformer.decoder.layers.3.catext_norm.bias": 256,
  "module.transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.3.norm2.weight": 256,
  "module.transformer.decoder.layers.3.norm2.bias": 256,
  "module.transformer.decoder.layers.3.linear1.weight": 524288,
  "module.transformer.decoder.layers.3.linear1.bias": 2048,
  "module.transformer.decoder.layers.3.linear2.weight": 524288,
  "module.transformer.decoder.layers.3.linear2.bias": 256,
  "module.transformer.decoder.layers.3.norm3.weight": 256,
  "module.transformer.decoder.layers.3.norm3.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm1.weight": 256,
  "module.transformer.decoder.layers.4.norm1.bias": 256,
  "module.transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.catext_norm.weight": 256,
  "module.transformer.decoder.layers.4.catext_norm.bias": 256,
  "module.transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.4.norm2.weight": 256,
  "module.transformer.decoder.layers.4.norm2.bias": 256,
  "module.transformer.decoder.layers.4.linear1.weight": 524288,
  "module.transformer.decoder.layers.4.linear1.bias": 2048,
  "module.transformer.decoder.layers.4.linear2.weight": 524288,
  "module.transformer.decoder.layers.4.linear2.bias": 256,
  "module.transformer.decoder.layers.4.norm3.weight": 256,
  "module.transformer.decoder.layers.4.norm3.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "module.transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "module.transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm1.weight": 256,
  "module.transformer.decoder.layers.5.norm1.bias": 256,
  "module.transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.catext_norm.weight": 256,
  "module.transformer.decoder.layers.5.catext_norm.bias": 256,
  "module.transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "module.transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "module.transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "module.transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "module.transformer.decoder.layers.5.norm2.weight": 256,
  "module.transformer.decoder.layers.5.norm2.bias": 256,
  "module.transformer.decoder.layers.5.linear1.weight": 524288,
  "module.transformer.decoder.layers.5.linear1.bias": 2048,
  "module.transformer.decoder.layers.5.linear2.weight": 524288,
  "module.transformer.decoder.layers.5.linear2.bias": 256,
  "module.transformer.decoder.layers.5.norm3.weight": 256,
  "module.transformer.decoder.layers.5.norm3.bias": 256,
  "module.transformer.decoder.norm.weight": 256,
  "module.transformer.decoder.norm.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "module.transformer.decoder.ref_point_head.layers.0.bias": 256,
  "module.transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "module.transformer.decoder.ref_point_head.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "module.transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "module.transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "module.transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "module.transformer.tgt_embed.weight": 230400,
  "module.transformer.enc_output.weight": 65536,
  "module.transformer.enc_output.bias": 256,
  "module.transformer.enc_output_norm.weight": 256,
  "module.transformer.enc_output_norm.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "module.transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "module.transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "module.transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "module.feat_map.weight": 196608,
  "module.feat_map.bias": 256,
  "module.input_proj.0.0.weight": 49152,
  "module.input_proj.0.0.bias": 256,
  "module.input_proj.0.1.weight": 256,
  "module.input_proj.0.1.bias": 256,
  "module.input_proj.1.0.weight": 98304,
  "module.input_proj.1.0.bias": 256,
  "module.input_proj.1.1.weight": 256,
  "module.input_proj.1.1.bias": 256,
  "module.input_proj.2.0.weight": 196608,
  "module.input_proj.2.0.bias": 256,
  "module.input_proj.2.1.weight": 256,
  "module.input_proj.2.1.bias": 256,
  "module.input_proj.3.0.weight": 1769472,
  "module.input_proj.3.0.bias": 256,
  "module.input_proj.3.1.weight": 256,
  "module.input_proj.3.1.bias": 256,
  "module.backbone.0.patch_embed.proj.weight": 4608,
  "module.backbone.0.patch_embed.proj.bias": 96,
  "module.backbone.0.patch_embed.norm.weight": 96,
  "module.backbone.0.patch_embed.norm.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "module.backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "module.backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "module.backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "module.backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "module.backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "module.backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "module.backbone.0.layers.0.downsample.reduction.weight": 73728,
  "module.backbone.0.layers.0.downsample.norm.weight": 384,
  "module.backbone.0.layers.0.downsample.norm.bias": 384,
  "module.backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "module.backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "module.backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "module.backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "module.backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "module.backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "module.backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "module.backbone.0.layers.1.downsample.reduction.weight": 294912,
  "module.backbone.0.layers.1.downsample.norm.weight": 768,
  "module.backbone.0.layers.1.downsample.norm.bias": 768,
  "module.backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "module.backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "module.backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "module.backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "module.backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "module.backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "module.backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "module.backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "module.backbone.0.layers.2.downsample.norm.weight": 1536,
  "module.backbone.0.layers.2.downsample.norm.bias": 1536,
  "module.backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "module.backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "module.backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "module.backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "module.backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "module.backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "module.backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "module.backbone.0.norm1.weight": 192,
  "module.backbone.0.norm1.bias": 192,
  "module.backbone.0.norm2.weight": 384,
  "module.backbone.0.norm2.bias": 384,
  "module.backbone.0.norm3.weight": 768,
  "module.backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-05 10:30:41,957 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-05 10:30:55,417 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-05 10:30:55,418 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-05 10:30:58,406 | [34mIgnore keys: [][0m
[32mINFO    [0m [32m2024-09-05 10:30:58,784 | [34m_IncompatibleKeys(missing_keys=[], unexpected_keys=['label_enc.weight', 'bert.embeddings.position_ids'])[0m
[32mINFO    [0m [32m2024-09-07 02:10:28,856 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 02:10:28,858 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 02:10:28,876 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 02:10:28,878 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 02:10:28,879 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:10:28,880 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:10:28,881 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 02:10:28,886 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:10:45,012 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 02:10:45,150 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 02:10:45,376 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 02:10:46,192 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 02:10:46,208 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:11:01,700 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 02:11:01,701 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 02:30:34,509 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 02:30:34,512 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 02:30:34,536 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 02:30:34,539 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 02:30:34,541 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:30:34,542 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:30:34,543 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 02:30:34,550 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:31:12,812 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 02:31:13,042 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 02:31:13,454 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 02:31:14,377 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 02:31:14,471 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:31:38,089 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 02:31:38,091 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 02:33:23,970 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 02:33:23,972 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 02:33:23,991 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 02:33:23,992 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 02:33:23,993 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:33:23,994 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:33:23,995 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 02:33:24,000 | [34mbuild model ... ...[0m
[32mINFO    [0m [32m2024-09-07 02:36:29,471 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 02:36:29,474 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 02:36:29,522 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 02:36:29,524 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 02:36:29,526 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:36:29,527 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:36:29,529 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 02:36:29,553 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:37:59,446 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 02:37:59,672 | [34mnumber of params:158915128[0m
[32mINFO    [0m [32m2024-09-07 02:37:59,855 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 02:38:00,766 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 02:38:00,784 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:38:17,320 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 02:38:17,322 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 02:45:53,820 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 02:45:53,823 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 02:45:53,842 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 02:45:53,843 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 02:45:53,844 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:45:53,845 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:45:53,847 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 02:45:53,852 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:48:42,740 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 02:48:42,889 | [34mnumber of params:144731272[0m
[32mINFO    [0m [32m2024-09-07 02:48:43,076 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 02:48:44,401 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 02:48:44,429 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:49:07,381 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 02:49:07,382 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 02:51:26,475 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 02:51:26,491 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 02:51:26,585 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 02:51:26,621 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 02:51:26,634 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:51:26,635 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:51:26,636 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 02:51:26,642 | [34mbuild model ... ...[0m
[32mINFO    [0m [32m2024-09-07 02:54:36,015 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 02:54:36,027 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 02:54:36,095 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 02:54:36,096 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 02:54:36,097 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:54:36,098 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 02:54:36,103 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 02:54:36,113 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:55:15,038 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 02:55:15,175 | [34mnumber of params:144731272[0m
[32mINFO    [0m [32m2024-09-07 02:55:15,502 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 02:55:16,766 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 02:55:16,804 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 02:55:39,778 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 02:55:39,779 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 03:27:22,268 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 03:27:22,270 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 03:27:22,289 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 03:27:22,291 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 03:27:22,292 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 03:27:22,292 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 03:27:22,294 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 03:27:22,299 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 03:29:24,424 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 03:29:24,630 | [34mnumber of params:144731272[0m
[32mINFO    [0m [32m2024-09-07 03:29:24,831 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 03:29:25,792 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 03:29:25,834 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 03:29:51,649 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 03:29:51,650 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 03:59:55,323 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 03:59:55,336 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 03:59:55,374 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 03:59:55,376 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 03:59:55,377 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 03:59:55,378 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 03:59:55,382 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 03:59:55,394 | [34mbuild model ... ...[0m
[32mINFO    [0m [32m2024-09-07 05:04:26,186 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 05:04:26,189 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 05:04:26,208 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 05:04:26,212 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 05:04:26,213 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:04:26,214 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:04:26,220 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 05:04:26,239 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 05:09:17,564 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 05:09:17,708 | [34mnumber of params:144731272[0m
[32mINFO    [0m [32m2024-09-07 05:09:17,891 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 05:09:18,533 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 05:09:18,550 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 05:09:34,246 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 05:09:34,247 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 05:38:06,027 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 05:38:06,029 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 05:38:06,048 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 05:38:06,050 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 05:38:06,051 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:38:06,052 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:38:06,053 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 05:38:06,059 | [34mbuild model ... ...[0m
[32mINFO    [0m [32m2024-09-07 05:38:50,689 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 05:38:50,691 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 05:38:50,710 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 05:38:50,711 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 05:38:50,712 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:38:50,713 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:38:50,714 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 05:38:50,719 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 05:39:38,239 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 05:39:38,377 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 05:39:38,578 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 05:39:39,288 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 05:39:39,303 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 05:39:55,291 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 05:39:55,293 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 05:43:56,811 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 05:43:56,813 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 05:43:56,834 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 05:43:56,836 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 05:43:56,837 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:43:56,838 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 05:43:56,839 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 05:43:56,845 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 05:59:22,737 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 05:59:22,903 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 05:59:23,117 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 05:59:23,894 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 05:59:23,936 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 05:59:40,347 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 05:59:40,348 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 06:13:07,430 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 06:13:07,432 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 06:13:07,452 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 06:13:07,454 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 06:13:07,455 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:13:07,456 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:13:07,457 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 06:13:07,462 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:13:52,230 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 06:13:52,381 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 06:13:52,578 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 06:14:32,007 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 06:14:32,029 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:14:48,042 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 06:14:48,043 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 06:15:36,252 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 06:15:36,255 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 06:15:36,274 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 06:15:36,275 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 06:15:36,276 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:15:36,277 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:15:36,278 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 06:15:36,284 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:16:02,508 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 06:16:02,645 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 06:16:02,845 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 06:16:03,524 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 06:16:03,542 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:16:19,626 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 06:16:19,627 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 06:18:34,595 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 06:18:34,598 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 06:18:34,616 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 06:18:34,618 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 06:18:34,619 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:18:34,620 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:18:34,621 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 06:18:34,626 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:18:54,295 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 06:18:54,438 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 06:18:54,634 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 06:18:55,321 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 06:18:55,338 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:19:11,967 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 06:19:11,968 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 06:23:31,799 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 06:23:31,801 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 06:23:31,821 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 06:23:31,823 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 06:23:31,824 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:23:31,825 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:23:31,826 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 06:23:31,832 | [34mbuild model ... ...[0m
[32mINFO    [0m [32m2024-09-07 06:26:48,628 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 06:26:48,631 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 06:26:48,651 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 06:26:48,652 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 06:26:48,653 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:26:48,654 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 06:26:48,655 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 06:26:48,660 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:27:17,693 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 06:27:17,828 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 06:27:18,030 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 06:27:18,802 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 06:27:18,819 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 06:27:35,919 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 06:27:35,920 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 10:07:49,419 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 10:07:49,421 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 10:07:49,439 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 10:07:49,441 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 10:07:49,442 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 10:07:49,442 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 10:07:49,444 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 10:07:49,448 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 10:10:33,533 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 10:10:33,664 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 10:10:33,847 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 10:10:34,511 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 10:10:34,526 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 10:10:48,090 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 10:10:48,092 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 15:27:00,021 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 15:27:00,024 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 15:27:00,043 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 15:27:00,045 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 15:27:00,046 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 15:27:00,047 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 15:27:00,048 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 15:27:00,053 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 15:27:34,959 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 15:27:35,107 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 15:27:35,309 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 15:27:36,002 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 15:27:36,019 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 15:27:54,110 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 15:27:54,113 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 15:29:58,501 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 15:29:58,504 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 15:29:58,523 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 15:29:58,525 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 15:29:58,526 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 15:29:58,526 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 15:29:58,528 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 15:29:58,534 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 15:30:24,053 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 15:30:24,186 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 15:30:24,392 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 15:30:25,057 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 15:30:25,073 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 15:30:41,990 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 15:30:41,991 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 16:18:53,538 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 16:18:53,540 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 16:18:53,559 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 16:18:53,561 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 16:18:53,561 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:18:53,562 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:18:53,564 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 16:18:53,569 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:20:57,247 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 16:20:57,382 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 16:20:57,568 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 16:20:58,213 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 16:20:58,229 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:21:14,476 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 16:21:14,478 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 16:28:11,015 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 16:28:11,018 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 16:28:11,037 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 16:28:11,038 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 16:28:11,039 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:28:11,040 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:28:11,041 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 16:28:11,047 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:28:34,757 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 16:28:34,907 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 16:28:35,096 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 16:28:35,770 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 16:28:35,785 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:28:52,379 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 16:28:52,380 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 16:43:07,604 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 16:43:07,606 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 16:43:07,626 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 16:43:07,627 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 16:43:07,628 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:43:07,629 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:43:07,630 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 16:43:07,635 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:43:25,639 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 16:43:25,783 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 16:43:25,994 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 16:43:26,683 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 16:43:26,700 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:43:43,497 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 16:43:43,499 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 16:51:23,419 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 16:51:23,421 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 16:51:23,441 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 16:51:23,443 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 16:51:23,444 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:51:23,444 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 16:51:23,446 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 16:51:23,451 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:51:42,042 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 16:51:42,179 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 16:51:42,372 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 16:51:43,060 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 16:51:43,077 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 16:51:59,995 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 16:51:59,997 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-07 17:07:01,381 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-07 17:07:01,383 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-07 17:07:01,402 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-07 17:07:01,404 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-07 17:07:01,405 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-07 17:07:01,406 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-07 17:07:01,407 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-07 17:07:01,412 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 17:07:16,574 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-07 17:07:16,718 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-07 17:07:16,922 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-07 17:07:17,610 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-07 17:07:17,627 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-07 17:07:34,520 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-07 17:07:34,521 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 06:58:10,346 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 06:58:10,349 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 06:58:10,368 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 06:58:10,371 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 06:58:10,372 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 06:58:10,372 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 06:58:10,374 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 06:58:10,379 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 06:58:29,464 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 06:58:29,607 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 06:58:29,816 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 06:58:30,474 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 06:58:30,495 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 06:58:48,901 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 06:58:48,903 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 07:22:39,907 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 07:22:39,909 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 07:22:39,928 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 07:22:39,930 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 07:22:39,930 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 07:22:39,931 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 07:22:39,932 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 07:22:39,937 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 07:22:53,294 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 07:22:53,419 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 07:22:53,610 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 07:22:54,265 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 07:22:54,280 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 07:23:08,042 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 07:23:08,043 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 07:41:15,637 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 07:41:15,639 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 07:41:15,658 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 07:41:15,660 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 07:41:15,661 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 07:41:15,662 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 07:41:15,663 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 07:41:15,668 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 07:41:29,942 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 07:41:30,079 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 07:41:30,277 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 07:41:30,955 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 07:41:30,972 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 07:41:47,565 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 07:41:47,567 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 07:49:39,707 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 07:49:39,709 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 07:49:39,728 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 07:49:39,729 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 07:49:39,730 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 07:49:39,731 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 07:49:39,732 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 07:49:39,737 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 07:49:52,621 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 07:49:52,754 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 07:49:52,947 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 07:49:53,611 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 07:49:53,626 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 07:50:07,838 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 07:50:07,840 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 08:01:30,793 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 08:01:30,795 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 08:01:30,814 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 08:01:30,816 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 08:01:30,816 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:01:30,817 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:01:30,818 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 08:01:30,823 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 08:01:44,395 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 08:01:44,529 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 08:01:44,724 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 08:01:45,388 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 08:01:45,404 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 08:01:59,433 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 08:01:59,434 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 08:09:18,946 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 08:09:18,949 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 08:09:18,968 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 08:09:18,969 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 08:09:18,970 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:09:18,971 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:09:18,973 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 08:09:18,978 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 08:09:31,680 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 08:09:31,806 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 08:09:31,990 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 08:09:32,653 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 08:09:32,669 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 08:09:46,204 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 08:09:46,205 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 08:31:07,356 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 08:31:07,358 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 08:31:07,377 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 08:31:07,378 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 08:31:07,379 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:31:07,380 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:31:07,381 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 08:31:07,386 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 08:31:20,904 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 08:31:21,041 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 08:31:21,225 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 08:31:21,887 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 08:31:21,902 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 08:31:36,240 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 08:31:36,241 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-08 08:40:06,992 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 08:40:06,994 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 08:40:07,012 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 08:40:07,013 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 08:40:07,014 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:40:07,015 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 08:40:07,016 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 08:40:07,021 | [34mbuild model ... ...[0m
[32mINFO    [0m [32m2024-09-08 12:16:38,826 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-08 12:16:38,828 | [34mCommand: /home/jiask/Open-GroundingDino-main/main.py --output_dir /home/jiask/Open-GroundingDino-main/output -c /home/jiask/Open-GroundingDino-main/config/cfg_odvg.py --datasets /home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-08 12:16:38,847 | [34mFull config saved to /home/jiask/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-08 12:16:38,849 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-08 12:16:38,850 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-08 12:16:38,850 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-08 12:16:38,852 | [34margs: Namespace(add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/home/jiask/mae-main/coco_dataset/annotations/instances_val2017.json', config_file='/home/jiask/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/home/jiask/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=8, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/home/jiask/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-08 12:16:38,857 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 12:16:53,509 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-08 12:16:53,647 | [34mnumber of params:172249090[0m
[32mINFO    [0m [32m2024-09-08 12:16:53,854 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-08 12:16:54,504 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-08 12:16:54,520 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-08 12:17:10,120 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-08 12:17:10,122 | [34mnumber of training dataset: 1, samples: 118287[0m
[32mINFO    [0m [32m2024-09-19 00:08:54,678 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:08:54,680 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:08:54,690 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:08:54,691 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:08:54,691 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:08:54,691 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:08:54,692 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:08:54,695 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:09:05,134 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:09:05,199 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:09:05,294 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:09:05,605 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:09:05,612 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:09:18,990 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:09:18,991 | [34mnumber of training dataset: 1, samples: 1[0m
[32mINFO    [0m [32m2024-09-19 00:09:43,121 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:09:43,122 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:09:43,131 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:09:43,131 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:09:43,132 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:09:43,132 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:09:43,132 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:09:43,134 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:09:48,469 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:09:48,534 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:09:48,628 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:09:48,932 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:09:48,939 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:10:02,783 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:10:02,784 | [34mnumber of training dataset: 1, samples: 1[0m
[32mINFO    [0m [32m2024-09-19 00:11:01,108 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:11:01,110 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:11:01,119 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:11:01,119 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:11:01,120 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:11:01,120 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:11:01,120 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:11:01,123 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:11:06,243 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:11:06,307 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:11:06,402 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:11:06,718 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:11:06,724 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:11:20,463 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:11:20,463 | [34mnumber of training dataset: 1, samples: 1[0m
[32mINFO    [0m [32m2024-09-19 00:19:59,610 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:19:59,612 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:19:59,622 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:19:59,623 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:19:59,623 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:19:59,624 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:19:59,624 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:19:59,627 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:20:06,484 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:20:06,548 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:20:06,641 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:20:06,954 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:20:06,960 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:20:20,322 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:20:20,322 | [34mnumber of training dataset: 1, samples: 1[0m
[32mINFO    [0m [32m2024-09-19 00:31:01,895 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:31:01,896 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:31:01,905 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:31:01,906 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:31:01,906 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:31:01,906 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:31:01,907 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:31:01,909 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:31:07,177 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:31:07,244 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:31:07,339 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:31:07,659 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:31:07,666 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:31:21,069 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:31:21,070 | [34mnumber of training dataset: 1, samples: 1[0m
[32mINFO    [0m [32m2024-09-19 00:35:47,346 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:35:47,347 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:35:47,358 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:35:47,359 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:35:47,359 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:35:47,360 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:35:47,360 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:35:47,363 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:35:52,509 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:35:52,574 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:35:52,670 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:35:52,989 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:35:52,996 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:36:06,372 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:36:06,373 | [34mnumber of training dataset: 1, samples: 1[0m
[32mINFO    [0m [32m2024-09-19 00:38:54,917 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:38:54,918 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:38:54,927 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:38:54,927 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:38:54,928 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:38:54,928 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:38:54,928 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:38:54,931 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:39:00,345 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:39:00,409 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:39:00,505 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:39:00,819 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:39:00,826 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:42:52,344 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:42:52,345 | [34mnumber of training dataset: 1, samples: 1[0m
[32mINFO    [0m [32m2024-09-19 00:43:19,880 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:43:19,882 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:43:19,892 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:43:19,893 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:43:19,893 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:43:19,894 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:43:19,894 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:43:19,897 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:43:25,507 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:43:25,572 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:43:25,666 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:43:25,978 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:43:25,985 | [34mbuild dataset ... ...[0m
[32mINFO    [0m [32m2024-09-19 00:52:54,103 | [34mgit:
  sha: N/A, status: clean, branch: N/A
[0m
[32mINFO    [0m [32m2024-09-19 00:52:54,105 | [34mCommand: /root/Open-GroundingDino-main/main.py --output_dir /root/Open-GroundingDino-main/output -c /root/Open-GroundingDino-main/config/cfg_odvg.py --datasets /root/Open-GroundingDino-main/config/datasets_mixed_odvg.json --options text_encoder_type=bert-base-uncased[0m
[32mINFO    [0m [32m2024-09-19 00:52:54,113 | [34mFull config saved to /root/Open-GroundingDino-main/output/config_args_all.json[0m
[32mINFO    [0m [32m2024-09-19 00:52:54,114 | [34mworld size: 1[0m
[32mINFO    [0m [32m2024-09-19 00:52:54,114 | [34mrank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:52:54,114 | [34mlocal_rank: 0[0m
[32mINFO    [0m [32m2024-09-19 00:52:54,115 | [34margs: Namespace(adaptive_mask='no_attn_mask', add_channel_attention=False, add_pos_value=False, amp=False, aux_loss=True, backbone='swin_T_224_1k', backbone_freeze_keywords=None, batch_norm_type='FrozenBatchNorm2d', batch_size=4, bbox_loss_coef=5.0, box_attn_type='roi_align', clip_max_norm=0.1, cls_loss_coef=2.0, coco_val_path='/root/annotations/instances_val2017.json', config_file='/root/Open-GroundingDino-main/config/cfg_odvg.py', dabdetr_deformable_decoder=False, dabdetr_deformable_encoder=False, dabdetr_yolo_like_anchor_update=False, data_aug_max_size=1333, data_aug_scale_overlap=None, data_aug_scales=[480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800], data_aug_scales2_crop=[384, 600], data_aug_scales2_resize=[400, 500, 600], datasets='/root/Open-GroundingDino-main/config/datasets_mixed_odvg.json', ddetr_lr_param=False, debug=False, dec_layer_number=None, dec_layers=6, dec_n_points=4, dec_pred_bbox_embed_share=True, dec_pred_class_embed_share=True, decoder_layer_noise=False, decoder_module_seq=['sa', 'ca', 'ffn'], decoder_sa_type='sa', device='cuda', dice_loss_coef=1.0, dilation=False, dim_feedforward=2048, dist_url='env://', distributed=False, dln_hw_noise=0.2, dln_xy_noise=0.2, dn_bbox_coef=1.0, dn_box_noise_scale=1.0, dn_label_coef=1.0, dn_label_noise_ratio=0.5, dn_labelbook_size=91, dn_scalar=100, dropout=0.0, ema_decay=0.9997, ema_epoch=0, embed_init_tgt=True, enc_layers=6, enc_loss_coef=1.0, enc_n_points=4, epochs=15, eval=False, find_unused_params=False, finetune_ignore=None, fix_refpoints_hw=-1, fix_size=False, focal_alpha=0.25, focal_gamma=2.0, freeze_keywords=['bert'], frozen_stages=2, frozen_weights=None, fusion_dropout=0.0, fusion_droppath=0.1, giou_loss_coef=2.0, hidden_dim=256, interm_loss_coef=1.0, local_rank=0, lr=0.0001, lr_backbone=1e-05, lr_backbone_names=['backbone.0', 'bert'], lr_drop=4, lr_drop_list=[4, 8], lr_linear_proj_mult=1e-05, lr_linear_proj_names=['ref_point_head', 'sampling_offsets'], mask_loss_coef=1.0, masks=False, match_unstable_error=True, matcher_type='HungarianMatcher', max_labels=50, max_text_len=256, modelname='groundingdino', multi_step_lr=False, nheads=8, nms_iou_threshold=-1, no_interm_box_loss=False, note='', num_feature_levels=4, num_patterns=0, num_queries=900, num_select=300, num_workers=2, onecyclelr=False, options={'text_encoder_type': 'bert-base-uncased'}, output_dir='/root/Open-GroundingDino-main/output', param_dict_type='ddetr_in_mmdet', pdetr3_bbox_embed_diff_each_layer=False, pdetr3_refHW=-1, pe_temperatureH=20, pe_temperatureW=20, position_embedding='sine', pre_norm=False, pretrain_model_path=None, query_dim=4, random_refpoints_xy=False, rank=0, remove_difficult=False, resume='', return_interm_indices=[1, 2, 3], save_checkpoint_interval=1, save_log=False, save_results=False, seed=42, set_cost_bbox=5.0, set_cost_class=1.0, set_cost_giou=2.0, start_epoch=0, sub_sentence_present=True, test=False, text_dropout=0.0, text_encoder_type='bert-base-uncased', transformer_activation='relu', two_stage_add_query_num=0, two_stage_bbox_embed_share=False, two_stage_class_embed_share=False, two_stage_default_hw=0.05, two_stage_keep_all_tokens=False, two_stage_learn_wh=False, two_stage_pat_embed=0, two_stage_type='standard', use_checkpoint=True, use_coco_eval=True, use_deformable_box_attn=False, use_detached_boxes_dec_out=False, use_ema=False, use_fusion_layer=True, use_text_cross_attention=True, use_text_enhancer=True, use_transformer_ckpt=True, weight_decay=0.0001, world_size=1)
[0m
[36mDEBUG   [0m [36m2024-09-19 00:52:54,117 | [34mbuild model ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:52:59,227 | [34mbuild model, done.[0m
[32mINFO    [0m [32m2024-09-19 00:52:59,291 | [34mnumber of params:172512258[0m
[32mINFO    [0m [32m2024-09-19 00:52:59,385 | [34mparams before freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "bert.embeddings.word_embeddings.weight": 23440896,
  "bert.embeddings.position_embeddings.weight": 393216,
  "bert.embeddings.token_type_embeddings.weight": 1536,
  "bert.embeddings.LayerNorm.weight": 768,
  "bert.embeddings.LayerNorm.bias": 768,
  "bert.encoder.layer.0.attention.self.query.weight": 589824,
  "bert.encoder.layer.0.attention.self.query.bias": 768,
  "bert.encoder.layer.0.attention.self.key.weight": 589824,
  "bert.encoder.layer.0.attention.self.key.bias": 768,
  "bert.encoder.layer.0.attention.self.value.weight": 589824,
  "bert.encoder.layer.0.attention.self.value.bias": 768,
  "bert.encoder.layer.0.attention.output.dense.weight": 589824,
  "bert.encoder.layer.0.attention.output.dense.bias": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.0.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.0.intermediate.dense.bias": 3072,
  "bert.encoder.layer.0.output.dense.weight": 2359296,
  "bert.encoder.layer.0.output.dense.bias": 768,
  "bert.encoder.layer.0.output.LayerNorm.weight": 768,
  "bert.encoder.layer.0.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.attention.self.query.weight": 589824,
  "bert.encoder.layer.1.attention.self.query.bias": 768,
  "bert.encoder.layer.1.attention.self.key.weight": 589824,
  "bert.encoder.layer.1.attention.self.key.bias": 768,
  "bert.encoder.layer.1.attention.self.value.weight": 589824,
  "bert.encoder.layer.1.attention.self.value.bias": 768,
  "bert.encoder.layer.1.attention.output.dense.weight": 589824,
  "bert.encoder.layer.1.attention.output.dense.bias": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.1.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.1.intermediate.dense.bias": 3072,
  "bert.encoder.layer.1.output.dense.weight": 2359296,
  "bert.encoder.layer.1.output.dense.bias": 768,
  "bert.encoder.layer.1.output.LayerNorm.weight": 768,
  "bert.encoder.layer.1.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.attention.self.query.weight": 589824,
  "bert.encoder.layer.2.attention.self.query.bias": 768,
  "bert.encoder.layer.2.attention.self.key.weight": 589824,
  "bert.encoder.layer.2.attention.self.key.bias": 768,
  "bert.encoder.layer.2.attention.self.value.weight": 589824,
  "bert.encoder.layer.2.attention.self.value.bias": 768,
  "bert.encoder.layer.2.attention.output.dense.weight": 589824,
  "bert.encoder.layer.2.attention.output.dense.bias": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.2.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.2.intermediate.dense.bias": 3072,
  "bert.encoder.layer.2.output.dense.weight": 2359296,
  "bert.encoder.layer.2.output.dense.bias": 768,
  "bert.encoder.layer.2.output.LayerNorm.weight": 768,
  "bert.encoder.layer.2.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.attention.self.query.weight": 589824,
  "bert.encoder.layer.3.attention.self.query.bias": 768,
  "bert.encoder.layer.3.attention.self.key.weight": 589824,
  "bert.encoder.layer.3.attention.self.key.bias": 768,
  "bert.encoder.layer.3.attention.self.value.weight": 589824,
  "bert.encoder.layer.3.attention.self.value.bias": 768,
  "bert.encoder.layer.3.attention.output.dense.weight": 589824,
  "bert.encoder.layer.3.attention.output.dense.bias": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.3.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.3.intermediate.dense.bias": 3072,
  "bert.encoder.layer.3.output.dense.weight": 2359296,
  "bert.encoder.layer.3.output.dense.bias": 768,
  "bert.encoder.layer.3.output.LayerNorm.weight": 768,
  "bert.encoder.layer.3.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.attention.self.query.weight": 589824,
  "bert.encoder.layer.4.attention.self.query.bias": 768,
  "bert.encoder.layer.4.attention.self.key.weight": 589824,
  "bert.encoder.layer.4.attention.self.key.bias": 768,
  "bert.encoder.layer.4.attention.self.value.weight": 589824,
  "bert.encoder.layer.4.attention.self.value.bias": 768,
  "bert.encoder.layer.4.attention.output.dense.weight": 589824,
  "bert.encoder.layer.4.attention.output.dense.bias": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.4.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.4.intermediate.dense.bias": 3072,
  "bert.encoder.layer.4.output.dense.weight": 2359296,
  "bert.encoder.layer.4.output.dense.bias": 768,
  "bert.encoder.layer.4.output.LayerNorm.weight": 768,
  "bert.encoder.layer.4.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.attention.self.query.weight": 589824,
  "bert.encoder.layer.5.attention.self.query.bias": 768,
  "bert.encoder.layer.5.attention.self.key.weight": 589824,
  "bert.encoder.layer.5.attention.self.key.bias": 768,
  "bert.encoder.layer.5.attention.self.value.weight": 589824,
  "bert.encoder.layer.5.attention.self.value.bias": 768,
  "bert.encoder.layer.5.attention.output.dense.weight": 589824,
  "bert.encoder.layer.5.attention.output.dense.bias": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.5.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.5.intermediate.dense.bias": 3072,
  "bert.encoder.layer.5.output.dense.weight": 2359296,
  "bert.encoder.layer.5.output.dense.bias": 768,
  "bert.encoder.layer.5.output.LayerNorm.weight": 768,
  "bert.encoder.layer.5.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.attention.self.query.weight": 589824,
  "bert.encoder.layer.6.attention.self.query.bias": 768,
  "bert.encoder.layer.6.attention.self.key.weight": 589824,
  "bert.encoder.layer.6.attention.self.key.bias": 768,
  "bert.encoder.layer.6.attention.self.value.weight": 589824,
  "bert.encoder.layer.6.attention.self.value.bias": 768,
  "bert.encoder.layer.6.attention.output.dense.weight": 589824,
  "bert.encoder.layer.6.attention.output.dense.bias": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.6.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.6.intermediate.dense.bias": 3072,
  "bert.encoder.layer.6.output.dense.weight": 2359296,
  "bert.encoder.layer.6.output.dense.bias": 768,
  "bert.encoder.layer.6.output.LayerNorm.weight": 768,
  "bert.encoder.layer.6.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.attention.self.query.weight": 589824,
  "bert.encoder.layer.7.attention.self.query.bias": 768,
  "bert.encoder.layer.7.attention.self.key.weight": 589824,
  "bert.encoder.layer.7.attention.self.key.bias": 768,
  "bert.encoder.layer.7.attention.self.value.weight": 589824,
  "bert.encoder.layer.7.attention.self.value.bias": 768,
  "bert.encoder.layer.7.attention.output.dense.weight": 589824,
  "bert.encoder.layer.7.attention.output.dense.bias": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.7.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.7.intermediate.dense.bias": 3072,
  "bert.encoder.layer.7.output.dense.weight": 2359296,
  "bert.encoder.layer.7.output.dense.bias": 768,
  "bert.encoder.layer.7.output.LayerNorm.weight": 768,
  "bert.encoder.layer.7.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.attention.self.query.weight": 589824,
  "bert.encoder.layer.8.attention.self.query.bias": 768,
  "bert.encoder.layer.8.attention.self.key.weight": 589824,
  "bert.encoder.layer.8.attention.self.key.bias": 768,
  "bert.encoder.layer.8.attention.self.value.weight": 589824,
  "bert.encoder.layer.8.attention.self.value.bias": 768,
  "bert.encoder.layer.8.attention.output.dense.weight": 589824,
  "bert.encoder.layer.8.attention.output.dense.bias": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.8.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.8.intermediate.dense.bias": 3072,
  "bert.encoder.layer.8.output.dense.weight": 2359296,
  "bert.encoder.layer.8.output.dense.bias": 768,
  "bert.encoder.layer.8.output.LayerNorm.weight": 768,
  "bert.encoder.layer.8.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.attention.self.query.weight": 589824,
  "bert.encoder.layer.9.attention.self.query.bias": 768,
  "bert.encoder.layer.9.attention.self.key.weight": 589824,
  "bert.encoder.layer.9.attention.self.key.bias": 768,
  "bert.encoder.layer.9.attention.self.value.weight": 589824,
  "bert.encoder.layer.9.attention.self.value.bias": 768,
  "bert.encoder.layer.9.attention.output.dense.weight": 589824,
  "bert.encoder.layer.9.attention.output.dense.bias": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.9.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.9.intermediate.dense.bias": 3072,
  "bert.encoder.layer.9.output.dense.weight": 2359296,
  "bert.encoder.layer.9.output.dense.bias": 768,
  "bert.encoder.layer.9.output.LayerNorm.weight": 768,
  "bert.encoder.layer.9.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.attention.self.query.weight": 589824,
  "bert.encoder.layer.10.attention.self.query.bias": 768,
  "bert.encoder.layer.10.attention.self.key.weight": 589824,
  "bert.encoder.layer.10.attention.self.key.bias": 768,
  "bert.encoder.layer.10.attention.self.value.weight": 589824,
  "bert.encoder.layer.10.attention.self.value.bias": 768,
  "bert.encoder.layer.10.attention.output.dense.weight": 589824,
  "bert.encoder.layer.10.attention.output.dense.bias": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.10.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.10.intermediate.dense.bias": 3072,
  "bert.encoder.layer.10.output.dense.weight": 2359296,
  "bert.encoder.layer.10.output.dense.bias": 768,
  "bert.encoder.layer.10.output.LayerNorm.weight": 768,
  "bert.encoder.layer.10.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.attention.self.query.weight": 589824,
  "bert.encoder.layer.11.attention.self.query.bias": 768,
  "bert.encoder.layer.11.attention.self.key.weight": 589824,
  "bert.encoder.layer.11.attention.self.key.bias": 768,
  "bert.encoder.layer.11.attention.self.value.weight": 589824,
  "bert.encoder.layer.11.attention.self.value.bias": 768,
  "bert.encoder.layer.11.attention.output.dense.weight": 589824,
  "bert.encoder.layer.11.attention.output.dense.bias": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.attention.output.LayerNorm.bias": 768,
  "bert.encoder.layer.11.intermediate.dense.weight": 2359296,
  "bert.encoder.layer.11.intermediate.dense.bias": 3072,
  "bert.encoder.layer.11.output.dense.weight": 2359296,
  "bert.encoder.layer.11.output.dense.bias": 768,
  "bert.encoder.layer.11.output.LayerNorm.weight": 768,
  "bert.encoder.layer.11.output.LayerNorm.bias": 768,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[32mINFO    [0m [32m2024-09-19 00:52:59,698 | [34mparams after freezing:
{
  "transformer.level_embed": 1024,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.0.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.0.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.0.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.0.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.0.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.0.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.0.norm1.weight": 256,
  "transformer.encoder.layers.0.norm1.bias": 256,
  "transformer.encoder.layers.0.linear1.weight": 524288,
  "transformer.encoder.layers.0.linear1.bias": 2048,
  "transformer.encoder.layers.0.linear2.weight": 524288,
  "transformer.encoder.layers.0.linear2.bias": 256,
  "transformer.encoder.layers.0.norm2.weight": 256,
  "transformer.encoder.layers.0.norm2.bias": 256,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.1.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.1.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.1.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.1.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.1.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.1.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.1.norm1.weight": 256,
  "transformer.encoder.layers.1.norm1.bias": 256,
  "transformer.encoder.layers.1.linear1.weight": 524288,
  "transformer.encoder.layers.1.linear1.bias": 2048,
  "transformer.encoder.layers.1.linear2.weight": 524288,
  "transformer.encoder.layers.1.linear2.bias": 256,
  "transformer.encoder.layers.1.norm2.weight": 256,
  "transformer.encoder.layers.1.norm2.bias": 256,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.2.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.2.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.2.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.2.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.2.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.2.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.2.norm1.weight": 256,
  "transformer.encoder.layers.2.norm1.bias": 256,
  "transformer.encoder.layers.2.linear1.weight": 524288,
  "transformer.encoder.layers.2.linear1.bias": 2048,
  "transformer.encoder.layers.2.linear2.weight": 524288,
  "transformer.encoder.layers.2.linear2.bias": 256,
  "transformer.encoder.layers.2.norm2.weight": 256,
  "transformer.encoder.layers.2.norm2.bias": 256,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.3.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.3.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.3.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.3.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.3.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.3.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.3.norm1.weight": 256,
  "transformer.encoder.layers.3.norm1.bias": 256,
  "transformer.encoder.layers.3.linear1.weight": 524288,
  "transformer.encoder.layers.3.linear1.bias": 2048,
  "transformer.encoder.layers.3.linear2.weight": 524288,
  "transformer.encoder.layers.3.linear2.bias": 256,
  "transformer.encoder.layers.3.norm2.weight": 256,
  "transformer.encoder.layers.3.norm2.bias": 256,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.4.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.4.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.4.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.4.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.4.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.4.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.4.norm1.weight": 256,
  "transformer.encoder.layers.4.norm1.bias": 256,
  "transformer.encoder.layers.4.linear1.weight": 524288,
  "transformer.encoder.layers.4.linear1.bias": 2048,
  "transformer.encoder.layers.4.linear2.weight": 524288,
  "transformer.encoder.layers.4.linear2.bias": 256,
  "transformer.encoder.layers.4.norm2.weight": 256,
  "transformer.encoder.layers.4.norm2.bias": 256,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.weight": 65536,
  "transformer.encoder.layers.5.self_attn.sampling_offsets.bias": 256,
  "transformer.encoder.layers.5.self_attn.attention_weights.weight": 32768,
  "transformer.encoder.layers.5.self_attn.attention_weights.bias": 128,
  "transformer.encoder.layers.5.self_attn.value_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.value_proj.bias": 256,
  "transformer.encoder.layers.5.self_attn.output_proj.weight": 65536,
  "transformer.encoder.layers.5.self_attn.output_proj.bias": 256,
  "transformer.encoder.layers.5.norm1.weight": 256,
  "transformer.encoder.layers.5.norm1.bias": 256,
  "transformer.encoder.layers.5.linear1.weight": 524288,
  "transformer.encoder.layers.5.linear1.bias": 2048,
  "transformer.encoder.layers.5.linear2.weight": 524288,
  "transformer.encoder.layers.5.linear2.bias": 256,
  "transformer.encoder.layers.5.norm2.weight": 256,
  "transformer.encoder.layers.5.norm2.bias": 256,
  "transformer.encoder.text_layers.0.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.0.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.0.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.0.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.0.linear1.weight": 262144,
  "transformer.encoder.text_layers.0.linear1.bias": 1024,
  "transformer.encoder.text_layers.0.linear2.weight": 262144,
  "transformer.encoder.text_layers.0.linear2.bias": 256,
  "transformer.encoder.text_layers.0.norm1.weight": 256,
  "transformer.encoder.text_layers.0.norm1.bias": 256,
  "transformer.encoder.text_layers.0.norm2.weight": 256,
  "transformer.encoder.text_layers.0.norm2.bias": 256,
  "transformer.encoder.text_layers.1.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.1.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.1.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.1.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.1.linear1.weight": 262144,
  "transformer.encoder.text_layers.1.linear1.bias": 1024,
  "transformer.encoder.text_layers.1.linear2.weight": 262144,
  "transformer.encoder.text_layers.1.linear2.bias": 256,
  "transformer.encoder.text_layers.1.norm1.weight": 256,
  "transformer.encoder.text_layers.1.norm1.bias": 256,
  "transformer.encoder.text_layers.1.norm2.weight": 256,
  "transformer.encoder.text_layers.1.norm2.bias": 256,
  "transformer.encoder.text_layers.2.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.2.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.2.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.2.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.2.linear1.weight": 262144,
  "transformer.encoder.text_layers.2.linear1.bias": 1024,
  "transformer.encoder.text_layers.2.linear2.weight": 262144,
  "transformer.encoder.text_layers.2.linear2.bias": 256,
  "transformer.encoder.text_layers.2.norm1.weight": 256,
  "transformer.encoder.text_layers.2.norm1.bias": 256,
  "transformer.encoder.text_layers.2.norm2.weight": 256,
  "transformer.encoder.text_layers.2.norm2.bias": 256,
  "transformer.encoder.text_layers.3.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.3.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.3.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.3.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.3.linear1.weight": 262144,
  "transformer.encoder.text_layers.3.linear1.bias": 1024,
  "transformer.encoder.text_layers.3.linear2.weight": 262144,
  "transformer.encoder.text_layers.3.linear2.bias": 256,
  "transformer.encoder.text_layers.3.norm1.weight": 256,
  "transformer.encoder.text_layers.3.norm1.bias": 256,
  "transformer.encoder.text_layers.3.norm2.weight": 256,
  "transformer.encoder.text_layers.3.norm2.bias": 256,
  "transformer.encoder.text_layers.4.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.4.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.4.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.4.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.4.linear1.weight": 262144,
  "transformer.encoder.text_layers.4.linear1.bias": 1024,
  "transformer.encoder.text_layers.4.linear2.weight": 262144,
  "transformer.encoder.text_layers.4.linear2.bias": 256,
  "transformer.encoder.text_layers.4.norm1.weight": 256,
  "transformer.encoder.text_layers.4.norm1.bias": 256,
  "transformer.encoder.text_layers.4.norm2.weight": 256,
  "transformer.encoder.text_layers.4.norm2.bias": 256,
  "transformer.encoder.text_layers.5.self_attn.in_proj_weight": 196608,
  "transformer.encoder.text_layers.5.self_attn.in_proj_bias": 768,
  "transformer.encoder.text_layers.5.self_attn.out_proj.weight": 65536,
  "transformer.encoder.text_layers.5.self_attn.out_proj.bias": 256,
  "transformer.encoder.text_layers.5.linear1.weight": 262144,
  "transformer.encoder.text_layers.5.linear1.bias": 1024,
  "transformer.encoder.text_layers.5.linear2.weight": 262144,
  "transformer.encoder.text_layers.5.linear2.bias": 256,
  "transformer.encoder.text_layers.5.norm1.weight": 256,
  "transformer.encoder.text_layers.5.norm1.bias": 256,
  "transformer.encoder.text_layers.5.norm2.weight": 256,
  "transformer.encoder.text_layers.5.norm2.bias": 256,
  "transformer.encoder.fusion_layers.0.gamma_v": 256,
  "transformer.encoder.fusion_layers.0.gamma_l": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.0.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.0.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.gamma_v": 256,
  "transformer.encoder.fusion_layers.1.gamma_l": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.1.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.1.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.gamma_v": 256,
  "transformer.encoder.fusion_layers.2.gamma_l": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.2.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.2.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.gamma_v": 256,
  "transformer.encoder.fusion_layers.3.gamma_l": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.3.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.3.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.gamma_v": 256,
  "transformer.encoder.fusion_layers.4.gamma_l": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.4.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.4.attn.out_l_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.gamma_v": 256,
  "transformer.encoder.fusion_layers.5.gamma_l": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_v.bias": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.weight": 256,
  "transformer.encoder.fusion_layers.5.layer_norm_l.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_v_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.values_l_proj.bias": 1024,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_v_proj.bias": 256,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.weight": 262144,
  "transformer.encoder.fusion_layers.5.attn.out_l_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.0.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.0.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.0.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.0.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.0.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.0.norm1.weight": 256,
  "transformer.decoder.layers.0.norm1.bias": 256,
  "transformer.decoder.layers.0.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.0.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.0.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.0.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.0.catext_norm.weight": 256,
  "transformer.decoder.layers.0.catext_norm.bias": 256,
  "transformer.decoder.layers.0.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.0.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.0.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.0.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.0.norm2.weight": 256,
  "transformer.decoder.layers.0.norm2.bias": 256,
  "transformer.decoder.layers.0.linear1.weight": 524288,
  "transformer.decoder.layers.0.linear1.bias": 2048,
  "transformer.decoder.layers.0.linear2.weight": 524288,
  "transformer.decoder.layers.0.linear2.bias": 256,
  "transformer.decoder.layers.0.norm3.weight": 256,
  "transformer.decoder.layers.0.norm3.bias": 256,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.1.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.1.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.1.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.1.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.1.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.1.norm1.weight": 256,
  "transformer.decoder.layers.1.norm1.bias": 256,
  "transformer.decoder.layers.1.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.1.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.1.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.1.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.1.catext_norm.weight": 256,
  "transformer.decoder.layers.1.catext_norm.bias": 256,
  "transformer.decoder.layers.1.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.1.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.1.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.1.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.1.norm2.weight": 256,
  "transformer.decoder.layers.1.norm2.bias": 256,
  "transformer.decoder.layers.1.linear1.weight": 524288,
  "transformer.decoder.layers.1.linear1.bias": 2048,
  "transformer.decoder.layers.1.linear2.weight": 524288,
  "transformer.decoder.layers.1.linear2.bias": 256,
  "transformer.decoder.layers.1.norm3.weight": 256,
  "transformer.decoder.layers.1.norm3.bias": 256,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.2.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.2.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.2.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.2.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.2.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.2.norm1.weight": 256,
  "transformer.decoder.layers.2.norm1.bias": 256,
  "transformer.decoder.layers.2.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.2.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.2.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.2.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.2.catext_norm.weight": 256,
  "transformer.decoder.layers.2.catext_norm.bias": 256,
  "transformer.decoder.layers.2.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.2.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.2.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.2.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.2.norm2.weight": 256,
  "transformer.decoder.layers.2.norm2.bias": 256,
  "transformer.decoder.layers.2.linear1.weight": 524288,
  "transformer.decoder.layers.2.linear1.bias": 2048,
  "transformer.decoder.layers.2.linear2.weight": 524288,
  "transformer.decoder.layers.2.linear2.bias": 256,
  "transformer.decoder.layers.2.norm3.weight": 256,
  "transformer.decoder.layers.2.norm3.bias": 256,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.3.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.3.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.3.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.3.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.3.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.3.norm1.weight": 256,
  "transformer.decoder.layers.3.norm1.bias": 256,
  "transformer.decoder.layers.3.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.3.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.3.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.3.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.3.catext_norm.weight": 256,
  "transformer.decoder.layers.3.catext_norm.bias": 256,
  "transformer.decoder.layers.3.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.3.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.3.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.3.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.3.norm2.weight": 256,
  "transformer.decoder.layers.3.norm2.bias": 256,
  "transformer.decoder.layers.3.linear1.weight": 524288,
  "transformer.decoder.layers.3.linear1.bias": 2048,
  "transformer.decoder.layers.3.linear2.weight": 524288,
  "transformer.decoder.layers.3.linear2.bias": 256,
  "transformer.decoder.layers.3.norm3.weight": 256,
  "transformer.decoder.layers.3.norm3.bias": 256,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.4.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.4.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.4.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.4.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.4.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.4.norm1.weight": 256,
  "transformer.decoder.layers.4.norm1.bias": 256,
  "transformer.decoder.layers.4.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.4.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.4.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.4.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.4.catext_norm.weight": 256,
  "transformer.decoder.layers.4.catext_norm.bias": 256,
  "transformer.decoder.layers.4.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.4.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.4.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.4.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.4.norm2.weight": 256,
  "transformer.decoder.layers.4.norm2.bias": 256,
  "transformer.decoder.layers.4.linear1.weight": 524288,
  "transformer.decoder.layers.4.linear1.bias": 2048,
  "transformer.decoder.layers.4.linear2.weight": 524288,
  "transformer.decoder.layers.4.linear2.bias": 256,
  "transformer.decoder.layers.4.norm3.weight": 256,
  "transformer.decoder.layers.4.norm3.bias": 256,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.sampling_offsets.bias": 256,
  "transformer.decoder.layers.5.cross_attn.attention_weights.weight": 32768,
  "transformer.decoder.layers.5.cross_attn.attention_weights.bias": 128,
  "transformer.decoder.layers.5.cross_attn.value_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.value_proj.bias": 256,
  "transformer.decoder.layers.5.cross_attn.output_proj.weight": 65536,
  "transformer.decoder.layers.5.cross_attn.output_proj.bias": 256,
  "transformer.decoder.layers.5.norm1.weight": 256,
  "transformer.decoder.layers.5.norm1.bias": 256,
  "transformer.decoder.layers.5.ca_text.in_proj_weight": 196608,
  "transformer.decoder.layers.5.ca_text.in_proj_bias": 768,
  "transformer.decoder.layers.5.ca_text.out_proj.weight": 65536,
  "transformer.decoder.layers.5.ca_text.out_proj.bias": 256,
  "transformer.decoder.layers.5.catext_norm.weight": 256,
  "transformer.decoder.layers.5.catext_norm.bias": 256,
  "transformer.decoder.layers.5.self_attn.in_proj_weight": 196608,
  "transformer.decoder.layers.5.self_attn.in_proj_bias": 768,
  "transformer.decoder.layers.5.self_attn.out_proj.weight": 65536,
  "transformer.decoder.layers.5.self_attn.out_proj.bias": 256,
  "transformer.decoder.layers.5.norm2.weight": 256,
  "transformer.decoder.layers.5.norm2.bias": 256,
  "transformer.decoder.layers.5.linear1.weight": 524288,
  "transformer.decoder.layers.5.linear1.bias": 2048,
  "transformer.decoder.layers.5.linear2.weight": 524288,
  "transformer.decoder.layers.5.linear2.bias": 256,
  "transformer.decoder.layers.5.norm3.weight": 256,
  "transformer.decoder.layers.5.norm3.bias": 256,
  "transformer.decoder.norm.weight": 256,
  "transformer.decoder.norm.bias": 256,
  "transformer.decoder.ref_point_head.layers.0.weight": 131072,
  "transformer.decoder.ref_point_head.layers.0.bias": 256,
  "transformer.decoder.ref_point_head.layers.1.weight": 65536,
  "transformer.decoder.ref_point_head.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.0.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.0.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.1.weight": 65536,
  "transformer.decoder.bbox_embed.0.layers.1.bias": 256,
  "transformer.decoder.bbox_embed.0.layers.2.weight": 1024,
  "transformer.decoder.bbox_embed.0.layers.2.bias": 4,
  "transformer.tgt_embed.weight": 230400,
  "transformer.enc_output.weight": 65536,
  "transformer.enc_output.bias": 256,
  "transformer.enc_output_norm.weight": 256,
  "transformer.enc_output_norm.bias": 256,
  "transformer.enc_out_bbox_embed.layers.0.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.0.bias": 256,
  "transformer.enc_out_bbox_embed.layers.1.weight": 65536,
  "transformer.enc_out_bbox_embed.layers.1.bias": 256,
  "transformer.enc_out_bbox_embed.layers.2.weight": 1024,
  "transformer.enc_out_bbox_embed.layers.2.bias": 4,
  "adaptivemask.encoders.0.linear.weight": 65536,
  "adaptivemask.encoders.0.linear.bias": 256,
  "adaptivemask.encoders.1.linear.weight": 65536,
  "adaptivemask.encoders.1.linear.bias": 256,
  "adaptivemask.encoders.2.linear.weight": 65536,
  "adaptivemask.encoders.2.linear.bias": 256,
  "adaptivemask.encoders.3.linear.weight": 65536,
  "adaptivemask.encoders.3.linear.bias": 256,
  "feat_map.weight": 196608,
  "feat_map.bias": 256,
  "input_proj.0.0.weight": 49152,
  "input_proj.0.0.bias": 256,
  "input_proj.0.1.weight": 256,
  "input_proj.0.1.bias": 256,
  "input_proj.1.0.weight": 98304,
  "input_proj.1.0.bias": 256,
  "input_proj.1.1.weight": 256,
  "input_proj.1.1.bias": 256,
  "input_proj.2.0.weight": 196608,
  "input_proj.2.0.bias": 256,
  "input_proj.2.1.weight": 256,
  "input_proj.2.1.bias": 256,
  "input_proj.3.0.weight": 1769472,
  "input_proj.3.0.bias": 256,
  "input_proj.3.1.weight": 256,
  "input_proj.3.1.bias": 256,
  "backbone.0.patch_embed.proj.weight": 4608,
  "backbone.0.patch_embed.proj.bias": 96,
  "backbone.0.patch_embed.norm.weight": 96,
  "backbone.0.patch_embed.norm.bias": 96,
  "backbone.0.layers.0.blocks.0.norm1.weight": 96,
  "backbone.0.layers.0.blocks.0.norm1.bias": 96,
  "backbone.0.layers.0.blocks.0.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.0.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.0.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.0.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.0.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.0.norm2.weight": 96,
  "backbone.0.layers.0.blocks.0.norm2.bias": 96,
  "backbone.0.layers.0.blocks.0.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.0.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.0.mlp.fc2.bias": 96,
  "backbone.0.layers.0.blocks.1.norm1.weight": 96,
  "backbone.0.layers.0.blocks.1.norm1.bias": 96,
  "backbone.0.layers.0.blocks.1.attn.relative_position_bias_table": 507,
  "backbone.0.layers.0.blocks.1.attn.qkv.weight": 27648,
  "backbone.0.layers.0.blocks.1.attn.qkv.bias": 288,
  "backbone.0.layers.0.blocks.1.attn.proj.weight": 9216,
  "backbone.0.layers.0.blocks.1.attn.proj.bias": 96,
  "backbone.0.layers.0.blocks.1.norm2.weight": 96,
  "backbone.0.layers.0.blocks.1.norm2.bias": 96,
  "backbone.0.layers.0.blocks.1.mlp.fc1.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc1.bias": 384,
  "backbone.0.layers.0.blocks.1.mlp.fc2.weight": 36864,
  "backbone.0.layers.0.blocks.1.mlp.fc2.bias": 96,
  "backbone.0.layers.0.downsample.reduction.weight": 73728,
  "backbone.0.layers.0.downsample.norm.weight": 384,
  "backbone.0.layers.0.downsample.norm.bias": 384,
  "backbone.0.layers.1.blocks.0.norm1.weight": 192,
  "backbone.0.layers.1.blocks.0.norm1.bias": 192,
  "backbone.0.layers.1.blocks.0.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.0.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.0.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.0.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.0.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.0.norm2.weight": 192,
  "backbone.0.layers.1.blocks.0.norm2.bias": 192,
  "backbone.0.layers.1.blocks.0.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.0.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.0.mlp.fc2.bias": 192,
  "backbone.0.layers.1.blocks.1.norm1.weight": 192,
  "backbone.0.layers.1.blocks.1.norm1.bias": 192,
  "backbone.0.layers.1.blocks.1.attn.relative_position_bias_table": 1014,
  "backbone.0.layers.1.blocks.1.attn.qkv.weight": 110592,
  "backbone.0.layers.1.blocks.1.attn.qkv.bias": 576,
  "backbone.0.layers.1.blocks.1.attn.proj.weight": 36864,
  "backbone.0.layers.1.blocks.1.attn.proj.bias": 192,
  "backbone.0.layers.1.blocks.1.norm2.weight": 192,
  "backbone.0.layers.1.blocks.1.norm2.bias": 192,
  "backbone.0.layers.1.blocks.1.mlp.fc1.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc1.bias": 768,
  "backbone.0.layers.1.blocks.1.mlp.fc2.weight": 147456,
  "backbone.0.layers.1.blocks.1.mlp.fc2.bias": 192,
  "backbone.0.layers.1.downsample.reduction.weight": 294912,
  "backbone.0.layers.1.downsample.norm.weight": 768,
  "backbone.0.layers.1.downsample.norm.bias": 768,
  "backbone.0.layers.2.blocks.0.norm1.weight": 384,
  "backbone.0.layers.2.blocks.0.norm1.bias": 384,
  "backbone.0.layers.2.blocks.0.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.0.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.0.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.0.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.0.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.0.norm2.weight": 384,
  "backbone.0.layers.2.blocks.0.norm2.bias": 384,
  "backbone.0.layers.2.blocks.0.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.0.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.0.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.1.norm1.weight": 384,
  "backbone.0.layers.2.blocks.1.norm1.bias": 384,
  "backbone.0.layers.2.blocks.1.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.1.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.1.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.1.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.1.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.1.norm2.weight": 384,
  "backbone.0.layers.2.blocks.1.norm2.bias": 384,
  "backbone.0.layers.2.blocks.1.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.1.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.1.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.2.norm1.weight": 384,
  "backbone.0.layers.2.blocks.2.norm1.bias": 384,
  "backbone.0.layers.2.blocks.2.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.2.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.2.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.2.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.2.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.2.norm2.weight": 384,
  "backbone.0.layers.2.blocks.2.norm2.bias": 384,
  "backbone.0.layers.2.blocks.2.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.2.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.2.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.3.norm1.weight": 384,
  "backbone.0.layers.2.blocks.3.norm1.bias": 384,
  "backbone.0.layers.2.blocks.3.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.3.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.3.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.3.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.3.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.3.norm2.weight": 384,
  "backbone.0.layers.2.blocks.3.norm2.bias": 384,
  "backbone.0.layers.2.blocks.3.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.3.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.3.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.4.norm1.weight": 384,
  "backbone.0.layers.2.blocks.4.norm1.bias": 384,
  "backbone.0.layers.2.blocks.4.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.4.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.4.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.4.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.4.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.4.norm2.weight": 384,
  "backbone.0.layers.2.blocks.4.norm2.bias": 384,
  "backbone.0.layers.2.blocks.4.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.4.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.4.mlp.fc2.bias": 384,
  "backbone.0.layers.2.blocks.5.norm1.weight": 384,
  "backbone.0.layers.2.blocks.5.norm1.bias": 384,
  "backbone.0.layers.2.blocks.5.attn.relative_position_bias_table": 2028,
  "backbone.0.layers.2.blocks.5.attn.qkv.weight": 442368,
  "backbone.0.layers.2.blocks.5.attn.qkv.bias": 1152,
  "backbone.0.layers.2.blocks.5.attn.proj.weight": 147456,
  "backbone.0.layers.2.blocks.5.attn.proj.bias": 384,
  "backbone.0.layers.2.blocks.5.norm2.weight": 384,
  "backbone.0.layers.2.blocks.5.norm2.bias": 384,
  "backbone.0.layers.2.blocks.5.mlp.fc1.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc1.bias": 1536,
  "backbone.0.layers.2.blocks.5.mlp.fc2.weight": 589824,
  "backbone.0.layers.2.blocks.5.mlp.fc2.bias": 384,
  "backbone.0.layers.2.downsample.reduction.weight": 1179648,
  "backbone.0.layers.2.downsample.norm.weight": 1536,
  "backbone.0.layers.2.downsample.norm.bias": 1536,
  "backbone.0.layers.3.blocks.0.norm1.weight": 768,
  "backbone.0.layers.3.blocks.0.norm1.bias": 768,
  "backbone.0.layers.3.blocks.0.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.0.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.0.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.0.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.0.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.0.norm2.weight": 768,
  "backbone.0.layers.3.blocks.0.norm2.bias": 768,
  "backbone.0.layers.3.blocks.0.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.0.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.0.mlp.fc2.bias": 768,
  "backbone.0.layers.3.blocks.1.norm1.weight": 768,
  "backbone.0.layers.3.blocks.1.norm1.bias": 768,
  "backbone.0.layers.3.blocks.1.attn.relative_position_bias_table": 4056,
  "backbone.0.layers.3.blocks.1.attn.qkv.weight": 1769472,
  "backbone.0.layers.3.blocks.1.attn.qkv.bias": 2304,
  "backbone.0.layers.3.blocks.1.attn.proj.weight": 589824,
  "backbone.0.layers.3.blocks.1.attn.proj.bias": 768,
  "backbone.0.layers.3.blocks.1.norm2.weight": 768,
  "backbone.0.layers.3.blocks.1.norm2.bias": 768,
  "backbone.0.layers.3.blocks.1.mlp.fc1.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc1.bias": 3072,
  "backbone.0.layers.3.blocks.1.mlp.fc2.weight": 2359296,
  "backbone.0.layers.3.blocks.1.mlp.fc2.bias": 768,
  "backbone.0.norm1.weight": 192,
  "backbone.0.norm1.bias": 192,
  "backbone.0.norm2.weight": 384,
  "backbone.0.norm2.bias": 384,
  "backbone.0.norm3.weight": 768,
  "backbone.0.norm3.bias": 768
}[0m
[36mDEBUG   [0m [36m2024-09-19 00:52:59,704 | [34mbuild dataset ... ...[0m
[36mDEBUG   [0m [36m2024-09-19 00:53:19,479 | [34mbuild dataset, done.[0m
[36mDEBUG   [0m [36m2024-09-19 00:53:19,480 | [34mnumber of training dataset: 1, samples: 118287[0m
