[2023-08-24 16:47:57,210 INFO] Use GPU: 0 for training
[2023-08-24 16:48:01,517 INFO] unlabeled data number: 100000, labeled data number 80
[2023-08-24 16:48:01,517 INFO] Create train and test data loaders
[2023-08-24 16:48:08,099 INFO] [!] data loader keys: dict_keys(['train_lb', 'train_ulb', 'eval', 'test'])
[2023-08-24 16:50:18,220 INFO] Create optimizer and scheduler
[2023-08-24 16:50:19,929 INFO] Number of Trainable Params: 110075908
[2023-08-24 16:50:20,486 INFO] Arguments: Namespace(save_dir='/liuzicheng/jwy/saved_models/usb_nlp', save_name='pseudolabel_ag_news_40_0', resume=True, load_path='/liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth', overwrite=True, use_tensorboard=True, use_wandb=False, use_aim=False, epoch=100, num_train_iter=102400, num_warmup_iter=5120, num_eval_iter=2048, num_log_iter=256, num_labels=80, batch_size=8, uratio=1, eval_batch_size=8, ema_m=0.9999, ulb_loss_ratio=1.0, optim='AdamW', lr=5e-05, momentum=0.9, weight_decay=0.0005, layer_decay=0.65, net='bert_base_uncased', net_from_name=False, use_pretrain=False, pretrain_path='', algorithm='pseudolabel', use_cat=False, amp=False, clip_grad=0, imb_algorithm=None, data_dir='/liuzicheng/data/data', dataset='ag_news', num_classes=4, train_sampler='RandomSampler', num_workers=4, include_lb_to_ulb=True, lb_imb_ratio=1, ulb_imb_ratio=1, ulb_num_labels=None, img_size=32, crop_ratio=0.875, max_length=512, max_length_seconds=4.0, sample_rate=16000, world_size=1, rank=0, dist_url='tcp://127.0.0.1:20667', dist_backend='nccl', seed=2, gpu=0, multiprocessing_distributed=False, c='/liuzicheng/jwy/otherbaseline-main2/config/usb_nlp/pseudolabel/pseudolabel_ag_news_40_0.yaml', p_cutoff=0.95, unsup_warm_up=0.4, clip=0.0, distributed=False, ulb_dest_len=100000, lb_dest_len=80)
[2023-08-24 16:50:20,487 INFO] Resume load path /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth does not exist
[2023-08-24 16:50:20,488 INFO] Model training
[2023-08-24 16:51:28,903 INFO] 256 iteration USE_EMA: True, train/sup_loss: 1.3443, train/unsup_loss: 0.0000, train/total_loss: 1.3443, train/util_ratio: 0.0000, train/run_time: 0.1627, lr: 0.0000, train/prefecth_time: 0.0037 
[2023-08-24 16:52:33,698 INFO] 512 iteration USE_EMA: True, train/sup_loss: 0.3694, train/unsup_loss: 0.0000, train/total_loss: 0.3694, train/util_ratio: 0.0000, train/run_time: 0.2949, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 16:53:39,084 INFO] 768 iteration USE_EMA: True, train/sup_loss: 0.0216, train/unsup_loss: 0.0207, train/total_loss: 0.0219, train/util_ratio: 0.7500, train/run_time: 0.2087, lr: 0.0000, train/prefecth_time: 0.0039 
[2023-08-24 16:54:46,579 INFO] 1024 iteration USE_EMA: True, train/sup_loss: 0.0086, train/unsup_loss: 0.0146, train/total_loss: 0.0090, train/util_ratio: 1.0000, train/run_time: 0.2302, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 16:55:52,095 INFO] 1280 iteration USE_EMA: True, train/sup_loss: 0.0037, train/unsup_loss: 0.0107, train/total_loss: 0.0040, train/util_ratio: 0.8750, train/run_time: 0.1702, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 16:56:54,546 INFO] 1536 iteration USE_EMA: True, train/sup_loss: 0.0023, train/unsup_loss: 0.0020, train/total_loss: 0.0023, train/util_ratio: 0.7500, train/run_time: 0.1421, lr: 0.0000, train/prefecth_time: 0.0022 
[2023-08-24 16:57:58,125 INFO] 1792 iteration USE_EMA: True, train/sup_loss: 0.0014, train/unsup_loss: 0.0022, train/total_loss: 0.0015, train/util_ratio: 1.0000, train/run_time: 0.2019, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 16:59:01,999 INFO] validating...
[2023-08-24 16:59:32,567 INFO] confusion matrix:
[[0.3384 0.006  0.0024 0.6532]
 [0.0496 0.214  0.0296 0.7068]
 [0.0412 0.     0.058  0.9008]
 [0.0108 0.     0.0088 0.9804]]
[2023-08-24 16:59:34,931 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 16:59:37,184 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 16:59:37,185 INFO] 2048 iteration, USE_EMA: True, train/sup_loss: 0.0009, train/unsup_loss: 0.0044, train/total_loss: 0.0011, train/util_ratio: 0.7500, train/run_time: 0.2652, eval/loss: 1.3352, eval/top-1-acc: 0.3977, eval/balanced_acc: 0.3977, eval/precision: 0.6578, eval/recall: 0.3977, eval/F1: 0.3472, lr: 0.0000, train/prefecth_time: 0.0025 BEST_EVAL_ACC: 0.3977, at 2048 iters
[2023-08-24 17:00:40,373 INFO] 2304 iteration USE_EMA: True, train/sup_loss: 0.0006, train/unsup_loss: 0.0073, train/total_loss: 0.0010, train/util_ratio: 1.0000, train/run_time: 0.2370, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 17:01:33,709 INFO] 2560 iteration USE_EMA: True, train/sup_loss: 0.0004, train/unsup_loss: 0.0075, train/total_loss: 0.0009, train/util_ratio: 1.0000, train/run_time: 0.1078, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-24 17:02:16,527 INFO] 2816 iteration USE_EMA: True, train/sup_loss: 0.0003, train/unsup_loss: 0.0003, train/total_loss: 0.0003, train/util_ratio: 0.8750, train/run_time: 0.1147, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-24 17:03:16,970 INFO] 3072 iteration USE_EMA: True, train/sup_loss: 0.0002, train/unsup_loss: 0.0017, train/total_loss: 0.0003, train/util_ratio: 0.8750, train/run_time: 0.1299, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-24 17:04:21,648 INFO] 3328 iteration USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0001, train/total_loss: 0.0001, train/util_ratio: 0.8750, train/run_time: 0.1593, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 17:05:24,911 INFO] 3584 iteration USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0012, train/total_loss: 0.0002, train/util_ratio: 0.7500, train/run_time: 0.2693, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:06:27,831 INFO] 3840 iteration USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0016, train/total_loss: 0.0002, train/util_ratio: 1.0000, train/run_time: 0.2079, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 17:07:31,688 INFO] validating...
[2023-08-24 17:08:01,929 INFO] confusion matrix:
[[8.972e-01 1.600e-02 3.120e-02 5.560e-02]
 [9.200e-02 7.880e-01 5.080e-02 6.920e-02]
 [2.092e-01 1.200e-03 5.948e-01 1.948e-01]
 [8.760e-02 8.000e-04 5.600e-02 8.556e-01]]
[2023-08-24 17:08:04,534 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 17:08:06,753 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 17:08:06,754 INFO] 4096 iteration, USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0005, train/total_loss: 0.0001, train/util_ratio: 0.8750, train/run_time: 0.3515, eval/loss: 1.1420, eval/top-1-acc: 0.7839, eval/balanced_acc: 0.7839, eval/precision: 0.8038, eval/recall: 0.7839, eval/F1: 0.7827, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.7839, at 4096 iters
[2023-08-24 17:09:09,727 INFO] 4352 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0002, train/total_loss: 0.0001, train/util_ratio: 0.8750, train/run_time: 0.2509, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 17:10:13,339 INFO] 4608 iteration USE_EMA: True, train/sup_loss: 0.0008, train/unsup_loss: 0.0003, train/total_loss: 0.0008, train/util_ratio: 0.8750, train/run_time: 0.1360, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-24 17:11:18,427 INFO] 4864 iteration USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0005, train/total_loss: 0.0002, train/util_ratio: 0.8750, train/run_time: 0.2327, lr: 0.0000, train/prefecth_time: 0.0130 
[2023-08-24 17:12:22,242 INFO] 5120 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0038, train/total_loss: 0.0005, train/util_ratio: 1.0000, train/run_time: 0.1476, lr: 0.0001, train/prefecth_time: 0.0024 
[2023-08-24 17:13:29,215 INFO] 5376 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1252, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-24 17:14:34,185 INFO] 5632 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0021, train/total_loss: 0.0003, train/util_ratio: 1.0000, train/run_time: 0.2937, lr: 0.0000, train/prefecth_time: 0.0044 
[2023-08-24 17:15:37,933 INFO] 5888 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.3310, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 17:16:40,879 INFO] validating...
[2023-08-24 17:16:58,269 INFO] confusion matrix:
[[9.192e-01 1.800e-02 4.240e-02 2.040e-02]
 [5.800e-02 8.976e-01 3.400e-02 1.040e-02]
 [1.504e-01 2.000e-03 7.484e-01 9.920e-02]
 [1.116e-01 8.000e-04 9.400e-02 7.936e-01]]
[2023-08-24 17:17:00,836 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 17:17:03,288 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 17:17:03,289 INFO] 6144 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0001, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1916, eval/loss: 0.7933, eval/top-1-acc: 0.8397, eval/balanced_acc: 0.8397, eval/precision: 0.8482, eval/recall: 0.8397, eval/F1: 0.8405, lr: 0.0000, train/prefecth_time: 0.0038 BEST_EVAL_ACC: 0.8397, at 6144 iters
[2023-08-24 17:17:45,093 INFO] 6400 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0020, train/total_loss: 0.0003, train/util_ratio: 1.0000, train/run_time: 0.1074, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 17:18:32,940 INFO] 6656 iteration USE_EMA: True, train/sup_loss: 0.2858, train/unsup_loss: 0.0058, train/total_loss: 0.2867, train/util_ratio: 0.8750, train/run_time: 0.3497, lr: 0.0000, train/prefecth_time: 0.0014 
[2023-08-24 17:19:21,645 INFO] 6912 iteration USE_EMA: True, train/sup_loss: 0.0002, train/unsup_loss: 0.0013, train/total_loss: 0.0004, train/util_ratio: 0.7500, train/run_time: 0.1131, lr: 0.0000, train/prefecth_time: 0.0026 
[2023-08-24 17:20:08,774 INFO] 7168 iteration USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0001, train/total_loss: 0.0002, train/util_ratio: 1.0000, train/run_time: 0.1608, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:21:14,056 INFO] 7424 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0001, train/total_loss: 0.0001, train/util_ratio: 1.0000, train/run_time: 0.3683, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:22:17,627 INFO] 7680 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0002, train/total_loss: 0.0001, train/util_ratio: 0.8750, train/run_time: 0.2623, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:23:21,523 INFO] 7936 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0001, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1518, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-24 17:24:26,342 INFO] validating...
[2023-08-24 17:24:56,988 INFO] confusion matrix:
[[9.216e-01 2.000e-02 4.360e-02 1.480e-02]
 [4.440e-02 9.240e-01 2.680e-02 4.800e-03]
 [1.272e-01 2.000e-03 7.800e-01 9.080e-02]
 [1.236e-01 8.000e-04 1.032e-01 7.724e-01]]
[2023-08-24 17:24:59,595 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 17:25:01,994 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 17:25:01,994 INFO] 8192 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0011, train/total_loss: 0.0002, train/util_ratio: 0.8750, train/run_time: 0.1358, eval/loss: 0.5447, eval/top-1-acc: 0.8495, eval/balanced_acc: 0.8495, eval/precision: 0.8566, eval/recall: 0.8495, eval/F1: 0.8499, lr: 0.0000, train/prefecth_time: 0.0032 BEST_EVAL_ACC: 0.8495, at 8192 iters
[2023-08-24 17:26:06,957 INFO] 8448 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2440, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:27:10,635 INFO] 8704 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0016, train/total_loss: 0.0004, train/util_ratio: 1.0000, train/run_time: 0.1333, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-24 17:28:15,354 INFO] 8960 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2279, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 17:29:20,120 INFO] 9216 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1795, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 17:30:26,131 INFO] 9472 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.2449, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 17:31:30,315 INFO] 9728 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3467, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:32:33,390 INFO] 9984 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0003, train/total_loss: 0.0001, train/util_ratio: 1.0000, train/run_time: 0.1343, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-24 17:33:37,323 INFO] validating...
[2023-08-24 17:34:07,093 INFO] confusion matrix:
[[0.9204 0.022  0.0456 0.012 ]
 [0.0328 0.9452 0.0184 0.0036]
 [0.1124 0.002  0.7968 0.0888]
 [0.1228 0.002  0.1052 0.77  ]]
[2023-08-24 17:34:09,792 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 17:34:12,708 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 17:34:12,709 INFO] 10240 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0001, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2613, eval/loss: 0.4571, eval/top-1-acc: 0.8581, eval/balanced_acc: 0.8581, eval/precision: 0.8633, eval/recall: 0.8581, eval/F1: 0.8581, lr: 0.0000, train/prefecth_time: 0.0025 BEST_EVAL_ACC: 0.8581, at 10240 iters
[2023-08-24 17:34:58,098 INFO] 10496 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1105, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-24 17:35:48,085 INFO] 10752 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2739, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 17:36:50,669 INFO] 11008 iteration USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0017, train/total_loss: 0.0005, train/util_ratio: 0.8750, train/run_time: 0.1537, lr: 0.0000, train/prefecth_time: 0.0023 
[2023-08-24 17:37:54,206 INFO] 11264 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0021, train/total_loss: 0.0006, train/util_ratio: 1.0000, train/run_time: 0.2605, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 17:38:58,122 INFO] 11520 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1253, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 17:40:02,102 INFO] 11776 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2057, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:41:08,914 INFO] 12032 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0020, train/total_loss: 0.0006, train/util_ratio: 1.0000, train/run_time: 0.3203, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-24 17:42:42,060 INFO] validating...
[2023-08-24 17:43:12,323 INFO] confusion matrix:
[[0.9156 0.0232 0.0504 0.0108]
 [0.0268 0.96   0.0116 0.0016]
 [0.1016 0.002  0.8168 0.0796]
 [0.1312 0.0024 0.1124 0.754 ]]
[2023-08-24 17:43:14,992 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 17:43:17,554 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 17:43:17,555 INFO] 12288 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.1872, eval/loss: 0.5267, eval/top-1-acc: 0.8616, eval/balanced_acc: 0.8616, eval/precision: 0.8666, eval/recall: 0.8616, eval/F1: 0.8613, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8616, at 12288 iters
[2023-08-24 17:44:54,047 INFO] 12544 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0002, train/total_loss: 0.0001, train/util_ratio: 1.0000, train/run_time: 0.3028, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 17:46:26,603 INFO] 12800 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2932, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 17:48:02,646 INFO] 13056 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0004, train/total_loss: 0.0001, train/util_ratio: 1.0000, train/run_time: 0.1728, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 17:49:27,643 INFO] 13312 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.1704, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-24 17:50:34,876 INFO] 13568 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1786, lr: 0.0000, train/prefecth_time: 0.0037 
[2023-08-24 17:51:35,769 INFO] 13824 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0024, train/total_loss: 0.0008, train/util_ratio: 1.0000, train/run_time: 0.3783, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 17:53:10,781 INFO] 14080 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0004, train/total_loss: 0.0001, train/util_ratio: 1.0000, train/run_time: 0.4774, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 17:54:45,142 INFO] validating...
[2023-08-24 17:55:14,741 INFO] confusion matrix:
[[0.9108 0.024  0.0568 0.0084]
 [0.0264 0.9604 0.0116 0.0016]
 [0.0888 0.002  0.8364 0.0728]
 [0.1384 0.004  0.1204 0.7372]]
[2023-08-24 17:55:17,138 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 17:55:17,139 INFO] 14336 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.4233, eval/loss: 0.6980, eval/top-1-acc: 0.8612, eval/balanced_acc: 0.8612, eval/precision: 0.8667, eval/recall: 0.8612, eval/F1: 0.8607, lr: 0.0000, train/prefecth_time: 0.0021 BEST_EVAL_ACC: 0.8616, at 12288 iters
[2023-08-24 17:56:52,742 INFO] 14592 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3747, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 17:58:27,593 INFO] 14848 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.2176, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 18:00:00,657 INFO] 15104 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2318, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 18:01:35,079 INFO] 15360 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1665, lr: 0.0000, train/prefecth_time: 0.0066 
[2023-08-24 18:03:10,950 INFO] 15616 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2736, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 18:04:45,753 INFO] 15872 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2870, lr: 0.0000, train/prefecth_time: 0.0022 
[2023-08-24 18:05:49,632 INFO] 16128 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0015, train/total_loss: 0.0006, train/util_ratio: 1.0000, train/run_time: 0.1611, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-24 18:06:48,842 INFO] validating...
[2023-08-24 18:07:19,925 INFO] confusion matrix:
[[0.9184 0.0252 0.0472 0.0092]
 [0.0192 0.9716 0.0076 0.0016]
 [0.1016 0.0032 0.8168 0.0784]
 [0.1372 0.0048 0.1064 0.7516]]
[2023-08-24 18:07:22,162 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 18:07:24,340 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 18:07:24,341 INFO] 16384 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3821, eval/loss: 0.9186, eval/top-1-acc: 0.8646, eval/balanced_acc: 0.8646, eval/precision: 0.8692, eval/recall: 0.8646, eval/F1: 0.8639, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8646, at 16384 iters
[2023-08-24 18:08:59,163 INFO] 16640 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.4279, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 18:10:35,438 INFO] 16896 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2956, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 18:12:08,759 INFO] 17152 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2104, lr: 0.0000, train/prefecth_time: 0.0004 
[2023-08-24 18:13:42,658 INFO] 17408 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2683, lr: 0.0000, train/prefecth_time: 0.0022 
[2023-08-24 18:15:13,492 INFO] 17664 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0046, train/total_loss: 0.0020, train/util_ratio: 1.0000, train/run_time: 0.2803, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-24 18:16:42,978 INFO] 17920 iteration USE_EMA: True, train/sup_loss: 0.0007, train/unsup_loss: 0.0031, train/total_loss: 0.0021, train/util_ratio: 0.8750, train/run_time: 0.2723, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-24 18:18:11,847 INFO] 18176 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3243, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-24 18:19:36,853 INFO] validating...
[2023-08-24 18:19:59,575 INFO] confusion matrix:
[[0.9228 0.0264 0.0412 0.0096]
 [0.0168 0.976  0.0052 0.002 ]
 [0.1256 0.0044 0.7852 0.0848]
 [0.13   0.0072 0.0912 0.7716]]
[2023-08-24 18:20:01,913 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 18:20:01,914 INFO] 18432 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.1538, eval/loss: 1.1884, eval/top-1-acc: 0.8639, eval/balanced_acc: 0.8639, eval/precision: 0.8686, eval/recall: 0.8639, eval/F1: 0.8632, lr: 0.0000, train/prefecth_time: 0.0038 BEST_EVAL_ACC: 0.8646, at 16384 iters
[2023-08-24 18:21:02,213 INFO] 18688 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0009, train/total_loss: 0.0004, train/util_ratio: 1.0000, train/run_time: 0.2207, lr: 0.0000, train/prefecth_time: 0.0035 
[2023-08-24 18:22:24,103 INFO] 18944 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2367, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 18:23:53,419 INFO] 19200 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.4824, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 18:25:23,036 INFO] 19456 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2425, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 18:26:55,515 INFO] 19712 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3015, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 18:28:26,513 INFO] 19968 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3037, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 18:31:31,739 INFO] 20224 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7603, lr: 0.0000, train/prefecth_time: 0.0042 
[2023-08-24 18:34:18,376 INFO] validating...
[2023-08-24 18:34:36,255 INFO] confusion matrix:
[[0.9188 0.0264 0.0436 0.0112]
 [0.0128 0.9796 0.0052 0.0024]
 [0.1152 0.0068 0.7824 0.0956]
 [0.1216 0.0084 0.0848 0.7852]]
[2023-08-24 18:34:38,628 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 18:34:40,998 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 18:34:40,999 INFO] 20480 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0013, train/total_loss: 0.0006, train/util_ratio: 1.0000, train/run_time: 0.3308, eval/loss: 1.3991, eval/top-1-acc: 0.8665, eval/balanced_acc: 0.8665, eval/precision: 0.8694, eval/recall: 0.8665, eval/F1: 0.8656, lr: 0.0000, train/prefecth_time: 0.0030 BEST_EVAL_ACC: 0.8665, at 20480 iters
[2023-08-24 18:37:25,073 INFO] 20736 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7633, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 18:40:22,829 INFO] 20992 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8111, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 18:43:23,470 INFO] 21248 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0014, train/total_loss: 0.0007, train/util_ratio: 1.0000, train/run_time: 0.6813, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 18:46:24,894 INFO] 21504 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6916, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 18:48:12,750 INFO] 21760 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2787, lr: 0.0000, train/prefecth_time: 0.0035 
[2023-08-24 18:50:59,413 INFO] 22016 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0016, train/total_loss: 0.0009, train/util_ratio: 1.0000, train/run_time: 0.7004, lr: 0.0000, train/prefecth_time: 0.0040 
[2023-08-24 18:54:10,720 INFO] 22272 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7547, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-24 18:57:24,114 INFO] validating...
[2023-08-24 18:57:53,399 INFO] confusion matrix:
[[0.918  0.0264 0.0444 0.0112]
 [0.014  0.9792 0.0044 0.0024]
 [0.1176 0.0076 0.7656 0.1092]
 [0.1192 0.008  0.0692 0.8036]]
[2023-08-24 18:57:55,667 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 18:57:57,857 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 18:57:57,858 INFO] 22528 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8994, eval/loss: 1.6701, eval/top-1-acc: 0.8666, eval/balanced_acc: 0.8666, eval/precision: 0.8695, eval/recall: 0.8666, eval/F1: 0.8657, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8666, at 22528 iters
[2023-08-24 19:01:02,826 INFO] 22784 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2879, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 19:02:56,114 INFO] 23040 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6845, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 19:05:51,575 INFO] 23296 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6487, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 19:08:49,839 INFO] 23552 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0006, train/total_loss: 0.0003, train/util_ratio: 1.0000, train/run_time: 0.6234, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 19:11:45,923 INFO] 23808 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8004, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 19:14:20,145 INFO] 24064 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3917, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 19:16:27,439 INFO] 24320 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7899, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 19:19:41,337 INFO] validating...
[2023-08-24 19:20:10,246 INFO] confusion matrix:
[[0.9092 0.0264 0.0496 0.0148]
 [0.0144 0.9788 0.0044 0.0024]
 [0.0888 0.0076 0.7948 0.1088]
 [0.1032 0.0084 0.0764 0.812 ]]
[2023-08-24 19:20:12,560 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 19:20:14,710 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 19:20:14,711 INFO] 24576 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.6673, eval/loss: 1.7851, eval/top-1-acc: 0.8737, eval/balanced_acc: 0.8737, eval/precision: 0.8745, eval/recall: 0.8737, eval/F1: 0.8729, lr: 0.0000, train/prefecth_time: 0.0023 BEST_EVAL_ACC: 0.8737, at 24576 iters
[2023-08-24 19:23:10,966 INFO] 24832 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5891, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 19:26:05,047 INFO] 25088 iteration USE_EMA: True, train/sup_loss: 0.0001, train/unsup_loss: 0.0005, train/total_loss: 0.0004, train/util_ratio: 1.0000, train/run_time: 0.5859, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 19:28:15,263 INFO] 25344 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2695, lr: 0.0000, train/prefecth_time: 0.0037 
[2023-08-24 19:30:53,001 INFO] 25600 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6397, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 19:33:50,329 INFO] 25856 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6481, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 19:36:46,589 INFO] 26112 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5825, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 19:39:55,755 INFO] 26368 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8366, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 19:41:46,642 INFO] validating...
[2023-08-24 19:42:07,583 INFO] confusion matrix:
[[0.906  0.0272 0.0512 0.0156]
 [0.0152 0.9776 0.0048 0.0024]
 [0.0844 0.0084 0.7996 0.1076]
 [0.1    0.0096 0.0772 0.8132]]
[2023-08-24 19:42:09,900 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 19:42:12,070 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/model_best.pth
[2023-08-24 19:42:12,071 INFO] 26624 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3115, eval/loss: 1.9706, eval/top-1-acc: 0.8741, eval/balanced_acc: 0.8741, eval/precision: 0.8747, eval/recall: 0.8741, eval/F1: 0.8734, lr: 0.0000, train/prefecth_time: 0.0036 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 19:45:15,653 INFO] 26880 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7171, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 19:48:14,800 INFO] 27136 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7369, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 19:51:10,740 INFO] 27392 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0007, train/total_loss: 0.0004, train/util_ratio: 1.0000, train/run_time: 0.5876, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 19:53:53,991 INFO] 27648 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3337, lr: 0.0000, train/prefecth_time: 0.0023 
[2023-08-24 19:56:00,546 INFO] 27904 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6132, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-24 19:58:56,905 INFO] 28160 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6278, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 20:02:06,138 INFO] 28416 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8094, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-24 20:05:33,121 INFO] validating...
[2023-08-24 20:06:02,306 INFO] confusion matrix:
[[0.9028 0.028  0.0512 0.018 ]
 [0.014  0.9792 0.004  0.0028]
 [0.0868 0.0088 0.7884 0.116 ]
 [0.0952 0.0096 0.0716 0.8236]]
[2023-08-24 20:06:04,680 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 20:06:04,681 INFO] 28672 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6392, eval/loss: 2.1598, eval/top-1-acc: 0.8735, eval/balanced_acc: 0.8735, eval/precision: 0.8738, eval/recall: 0.8735, eval/F1: 0.8727, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 20:08:14,531 INFO] 28928 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2999, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-24 20:10:52,299 INFO] 29184 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.7211, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 20:13:57,749 INFO] 29440 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6396, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 20:16:58,816 INFO] 29696 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5989, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 20:20:00,779 INFO] 29952 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6837, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 20:21:52,551 INFO] 30208 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6433, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 20:25:03,544 INFO] 30464 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8062, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-24 20:28:17,537 INFO] validating...
[2023-08-24 20:28:46,517 INFO] confusion matrix:
[[0.9008 0.028  0.0536 0.0176]
 [0.014  0.9788 0.0044 0.0028]
 [0.0816 0.0096 0.7964 0.1124]
 [0.0932 0.0116 0.078  0.8172]]
[2023-08-24 20:28:48,873 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 20:28:48,874 INFO] 30720 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6725, eval/loss: 2.2914, eval/top-1-acc: 0.8733, eval/balanced_acc: 0.8733, eval/precision: 0.8733, eval/recall: 0.8733, eval/F1: 0.8725, lr: 0.0000, train/prefecth_time: 0.0025 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 20:31:45,599 INFO] 30976 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7420, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-24 20:34:18,961 INFO] 31232 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0037, train/total_loss: 0.0028, train/util_ratio: 1.0000, train/run_time: 0.3080, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-24 20:36:12,650 INFO] 31488 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5868, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 20:39:14,364 INFO] 31744 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6054, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 20:42:16,130 INFO] 32000 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6731, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-24 20:45:11,185 INFO] 32256 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6142, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 20:47:39,212 INFO] 32512 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3143, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 20:50:14,603 INFO] validating...
[2023-08-24 20:50:44,209 INFO] confusion matrix:
[[0.9052 0.0272 0.0548 0.0128]
 [0.016  0.9772 0.0044 0.0024]
 [0.0836 0.0096 0.8012 0.1056]
 [0.1128 0.0112 0.0856 0.7904]]
[2023-08-24 20:50:46,548 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 20:50:46,549 INFO] 32768 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0005, train/total_loss: 0.0004, train/util_ratio: 1.0000, train/run_time: 0.6072, eval/loss: 2.4500, eval/top-1-acc: 0.8685, eval/balanced_acc: 0.8685, eval/precision: 0.8694, eval/recall: 0.8685, eval/F1: 0.8676, lr: 0.0000, train/prefecth_time: 0.0025 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 20:53:42,067 INFO] 33024 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0010, train/total_loss: 0.0008, train/util_ratio: 1.0000, train/run_time: 0.6917, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 20:56:39,772 INFO] 33280 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7437, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 20:59:37,980 INFO] 33536 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6426, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 21:01:25,484 INFO] 33792 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2626, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 21:04:09,528 INFO] 34048 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7448, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-24 21:07:13,401 INFO] 34304 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6038, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-24 21:10:19,595 INFO] 34560 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7968, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 21:13:40,123 INFO] validating...
[2023-08-24 21:14:08,974 INFO] confusion matrix:
[[0.896  0.0268 0.0656 0.0116]
 [0.0176 0.976  0.0048 0.0016]
 [0.0712 0.0092 0.8252 0.0944]
 [0.1164 0.0124 0.1064 0.7648]]
[2023-08-24 21:14:11,621 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 21:14:11,622 INFO] 34816 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6011, eval/loss: 2.5010, eval/top-1-acc: 0.8655, eval/balanced_acc: 0.8655, eval/precision: 0.8667, eval/recall: 0.8655, eval/F1: 0.8646, lr: 0.0000, train/prefecth_time: 0.0045 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 21:16:07,333 INFO] 35072 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6348, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-24 21:19:04,062 INFO] 35328 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6303, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 21:21:59,515 INFO] 35584 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0008, train/total_loss: 0.0007, train/util_ratio: 1.0000, train/run_time: 0.6698, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 21:24:57,625 INFO] 35840 iteration USE_EMA: True, train/sup_loss: 0.0005, train/unsup_loss: 0.0001, train/total_loss: 0.0006, train/util_ratio: 1.0000, train/run_time: 0.6893, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-24 21:27:11,886 INFO] 36096 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3004, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 21:29:10,967 INFO] 36352 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0020, train/total_loss: 0.0018, train/util_ratio: 1.0000, train/run_time: 0.7414, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 21:32:17,339 INFO] 36608 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8393, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-24 21:35:38,149 INFO] validating...
[2023-08-24 21:36:06,750 INFO] confusion matrix:
[[0.8972 0.0268 0.0644 0.0116]
 [0.0172 0.9764 0.0048 0.0016]
 [0.0716 0.01   0.8344 0.084 ]
 [0.12   0.014  0.1252 0.7408]]
[2023-08-24 21:36:09,058 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 21:36:09,058 INFO] 36864 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5936, eval/loss: 2.5292, eval/top-1-acc: 0.8622, eval/balanced_acc: 0.8622, eval/precision: 0.8642, eval/recall: 0.8622, eval/F1: 0.8610, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 21:39:05,005 INFO] 37120 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6378, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 21:41:29,513 INFO] 37376 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3033, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-24 21:43:57,628 INFO] 37632 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6831, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 21:46:54,091 INFO] 37888 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6323, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 21:49:51,040 INFO] 38144 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6378, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 21:52:45,229 INFO] 38400 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6026, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 21:54:42,126 INFO] 38656 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3385, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 21:57:39,272 INFO] validating...
[2023-08-24 21:58:09,103 INFO] confusion matrix:
[[0.902  0.0268 0.0596 0.0116]
 [0.0168 0.9768 0.0048 0.0016]
 [0.0808 0.0112 0.8204 0.0876]
 [0.1292 0.0148 0.1148 0.7412]]
[2023-08-24 21:58:11,532 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 21:58:11,532 INFO] 38912 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7419, eval/loss: 2.6292, eval/top-1-acc: 0.8601, eval/balanced_acc: 0.8601, eval/precision: 0.8622, eval/recall: 0.8601, eval/F1: 0.8588, lr: 0.0000, train/prefecth_time: 0.0046 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 22:01:16,760 INFO] 39168 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7601, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 22:04:11,967 INFO] 39424 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6299, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:07:08,504 INFO] 39680 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6065, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:08:40,466 INFO] 39936 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2601, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 22:10:13,027 INFO] 40192 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6399, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:13:07,492 INFO] 40448 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6476, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 22:16:05,027 INFO] 40704 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5647, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 22:19:31,855 INFO] validating...
[2023-08-24 22:20:02,592 INFO] confusion matrix:
[[0.9024 0.0276 0.0608 0.0092]
 [0.0156 0.9792 0.004  0.0012]
 [0.0792 0.014  0.8296 0.0772]
 [0.1344 0.0172 0.1364 0.712 ]]
[2023-08-24 22:20:05,066 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 22:20:05,067 INFO] 40960 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.8187, eval/loss: 2.7979, eval/top-1-acc: 0.8558, eval/balanced_acc: 0.8558, eval/precision: 0.8590, eval/recall: 0.8558, eval/F1: 0.8540, lr: 0.0000, train/prefecth_time: 0.0040 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 22:22:44,695 INFO] 41216 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8495, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:24:37,958 INFO] 41472 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6179, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:27:38,806 INFO] 41728 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9735, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-24 22:30:32,865 INFO] 41984 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7155, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:33:30,457 INFO] 42240 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5898, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:36:01,474 INFO] 42496 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.4453, lr: 0.0000, train/prefecth_time: 0.0037 
[2023-08-24 22:38:19,697 INFO] 42752 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0005, train/total_loss: 0.0005, train/util_ratio: 1.0000, train/run_time: 0.6080, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:41:40,293 INFO] validating...
[2023-08-24 22:42:11,620 INFO] confusion matrix:
[[0.8984 0.028  0.0632 0.0104]
 [0.0144 0.98   0.0044 0.0012]
 [0.0772 0.0116 0.8328 0.0784]
 [0.1244 0.018  0.1424 0.7152]]
[2023-08-24 22:42:14,047 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 22:42:14,048 INFO] 43008 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8271, eval/loss: 2.8996, eval/top-1-acc: 0.8566, eval/balanced_acc: 0.8566, eval/precision: 0.8594, eval/recall: 0.8566, eval/F1: 0.8549, lr: 0.0000, train/prefecth_time: 0.0030 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 22:45:13,422 INFO] 43264 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6836, lr: 0.0000, train/prefecth_time: 0.0041 
[2023-08-24 22:48:09,379 INFO] 43520 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6757, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 22:50:08,414 INFO] 43776 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3674, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-24 22:52:52,204 INFO] 44032 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6523, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 22:55:49,775 INFO] 44288 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6309, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 22:58:42,951 INFO] 44544 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6197, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-24 23:01:37,585 INFO] 44800 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6071, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 23:03:32,919 INFO] validating...
[2023-08-24 23:04:02,850 INFO] confusion matrix:
[[0.902  0.028  0.0596 0.0104]
 [0.0156 0.9788 0.0044 0.0012]
 [0.0892 0.0112 0.8124 0.0872]
 [0.1252 0.0156 0.128  0.7312]]
[2023-08-24 23:04:05,544 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 23:04:05,545 INFO] 45056 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.4592, eval/loss: 3.0705, eval/top-1-acc: 0.8561, eval/balanced_acc: 0.8561, eval/precision: 0.8584, eval/recall: 0.8561, eval/F1: 0.8546, lr: 0.0000, train/prefecth_time: 0.0087 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 23:12:12,392 INFO] 45312 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5940, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 23:15:04,212 INFO] 45568 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6495, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 23:17:59,961 INFO] 45824 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6316, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-24 23:20:51,563 INFO] 46080 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2751, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-24 23:22:18,228 INFO] 46336 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3002, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 23:23:58,102 INFO] 46592 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2985, lr: 0.0000, train/prefecth_time: 0.0044 
[2023-08-24 23:26:51,680 INFO] 46848 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6808, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 23:30:00,239 INFO] validating...
[2023-08-24 23:30:30,928 INFO] confusion matrix:
[[0.9092 0.0276 0.0532 0.01  ]
 [0.016  0.9788 0.004  0.0012]
 [0.108  0.0128 0.7792 0.1   ]
 [0.1308 0.0156 0.108  0.7456]]
[2023-08-24 23:30:33,265 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 23:30:33,266 INFO] 47104 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8118, eval/loss: 3.3854, eval/top-1-acc: 0.8532, eval/balanced_acc: 0.8532, eval/precision: 0.8556, eval/recall: 0.8532, eval/F1: 0.8517, lr: 0.0000, train/prefecth_time: 0.0025 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 23:33:47,442 INFO] 47360 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6091, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 23:36:29,465 INFO] 47616 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2946, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-24 23:38:34,774 INFO] 47872 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7974, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 23:41:29,871 INFO] 48128 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6204, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-24 23:44:28,113 INFO] 48384 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5965, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-24 23:47:26,395 INFO] 48640 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.4386, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-24 23:53:51,577 INFO] 48896 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2781, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-24 23:55:12,264 INFO] validating...
[2023-08-24 23:55:36,053 INFO] confusion matrix:
[[0.9108 0.0288 0.0508 0.0096]
 [0.0128 0.9828 0.0032 0.0012]
 [0.1176 0.0172 0.7592 0.106 ]
 [0.1372 0.018  0.0976 0.7472]]
[2023-08-24 23:55:38,468 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-24 23:55:38,468 INFO] 49152 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2731, eval/loss: 3.7397, eval/top-1-acc: 0.8500, eval/balanced_acc: 0.8500, eval/precision: 0.8525, eval/recall: 0.8500, eval/F1: 0.8482, lr: 0.0000, train/prefecth_time: 0.0027 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-24 23:57:34,124 INFO] 49408 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8151, lr: 0.0000, train/prefecth_time: 0.0026 
[2023-08-25 00:00:33,410 INFO] 49664 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6352, lr: 0.0000, train/prefecth_time: 0.0020 
[2023-08-25 00:03:25,738 INFO] 49920 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0047, train/total_loss: 0.0047, train/util_ratio: 1.0000, train/run_time: 0.8701, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 00:06:17,666 INFO] 50176 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7139, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 00:08:48,446 INFO] 50432 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0002, train/total_loss: 0.0002, train/util_ratio: 1.0000, train/run_time: 0.2691, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-25 00:11:04,229 INFO] 50688 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5837, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 00:13:57,121 INFO] 50944 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7473, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 00:16:52,696 INFO] validating...
[2023-08-25 00:17:22,213 INFO] confusion matrix:
[[0.9112 0.0296 0.05   0.0092]
 [0.0108 0.9848 0.0032 0.0012]
 [0.1248 0.02   0.7472 0.108 ]
 [0.1472 0.0232 0.0996 0.73  ]]
[2023-08-25 00:17:24,558 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 00:17:24,559 INFO] 51200 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6299, eval/loss: 4.1377, eval/top-1-acc: 0.8433, eval/balanced_acc: 0.8433, eval/precision: 0.8462, eval/recall: 0.8433, eval/F1: 0.8411, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 00:20:47,352 INFO] 51456 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8309, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 00:22:56,177 INFO] 51712 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2562, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-25 00:25:28,588 INFO] 51968 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7113, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-25 00:28:26,979 INFO] 52224 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6140, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 00:31:37,131 INFO] 52480 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 2.0490, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 00:37:40,014 INFO] 52736 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6453, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 00:39:49,530 INFO] 52992 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2726, lr: 0.0000, train/prefecth_time: 0.0035 
[2023-08-25 00:41:15,636 INFO] validating...
[2023-08-25 00:41:33,261 INFO] confusion matrix:
[[0.9136 0.0308 0.0452 0.0104]
 [0.0112 0.984  0.0036 0.0012]
 [0.1392 0.0228 0.7272 0.1108]
 [0.154  0.0248 0.0972 0.724 ]]
[2023-08-25 00:41:35,910 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 00:41:35,914 INFO] 53248 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2732, eval/loss: 4.5297, eval/top-1-acc: 0.8372, eval/balanced_acc: 0.8372, eval/precision: 0.8411, eval/recall: 0.8372, eval/F1: 0.8347, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 00:44:34,459 INFO] 53504 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7768, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 00:47:44,887 INFO] 53760 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7299, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 00:50:50,307 INFO] 54016 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6431, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 00:53:51,179 INFO] 54272 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6874, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 00:55:42,605 INFO] 54528 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2545, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-25 00:58:23,261 INFO] 54784 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6360, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:01:21,201 INFO] 55040 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6437, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 01:04:18,313 INFO] validating...
[2023-08-25 01:04:47,502 INFO] confusion matrix:
[[0.9172 0.0312 0.0416 0.01  ]
 [0.012  0.9836 0.0032 0.0012]
 [0.1532 0.028  0.7084 0.1104]
 [0.1596 0.0252 0.0948 0.7204]]
[2023-08-25 01:04:49,846 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 01:04:49,847 INFO] 55296 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6811, eval/loss: 4.8532, eval/top-1-acc: 0.8324, eval/balanced_acc: 0.8324, eval/precision: 0.8376, eval/recall: 0.8324, eval/F1: 0.8296, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 01:07:52,451 INFO] 55552 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2950, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-25 01:09:57,565 INFO] 55808 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0001, train/total_loss: 0.0001, train/util_ratio: 1.0000, train/run_time: 0.7080, lr: 0.0000, train/prefecth_time: 0.0067 
[2023-08-25 01:12:55,404 INFO] 56064 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7707, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:15:56,084 INFO] 56320 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6960, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:19:00,547 INFO] 56576 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6517, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:21:23,111 INFO] 56832 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2914, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-25 01:23:33,872 INFO] 57088 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6609, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 01:26:30,983 INFO] validating...
[2023-08-25 01:27:01,437 INFO] confusion matrix:
[[0.92   0.0304 0.04   0.0096]
 [0.0132 0.9824 0.0032 0.0012]
 [0.1648 0.0268 0.7012 0.1072]
 [0.1788 0.022  0.0952 0.704 ]]
[2023-08-25 01:27:03,825 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 01:27:03,826 INFO] 57344 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6271, eval/loss: 4.9310, eval/top-1-acc: 0.8269, eval/balanced_acc: 0.8269, eval/precision: 0.8344, eval/recall: 0.8269, eval/F1: 0.8241, lr: 0.0000, train/prefecth_time: 0.0035 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 01:30:14,125 INFO] 57600 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7756, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-25 01:33:33,611 INFO] 57856 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6527, lr: 0.0000, train/prefecth_time: 0.0062 
[2023-08-25 01:35:29,743 INFO] 58112 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7503, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:38:29,133 INFO] 58368 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6084, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:41:29,695 INFO] 58624 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7214, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:44:36,458 INFO] 58880 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.5837, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 01:47:17,199 INFO] 59136 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.4188, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-25 01:49:04,411 INFO] validating...
[2023-08-25 01:49:33,473 INFO] confusion matrix:
[[9.244e-01 2.880e-02 3.840e-02 8.400e-03]
 [1.640e-02 9.800e-01 2.800e-03 8.000e-04]
 [1.760e-01 2.640e-02 7.008e-01 9.680e-02]
 [1.924e-01 2.000e-02 1.056e-01 6.820e-01]]
[2023-08-25 01:49:36,162 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 01:49:36,163 INFO] 59392 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6755, eval/loss: 5.0276, eval/top-1-acc: 0.8218, eval/balanced_acc: 0.8218, eval/precision: 0.8318, eval/recall: 0.8218, eval/F1: 0.8189, lr: 0.0000, train/prefecth_time: 0.0046 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 01:52:42,850 INFO] 59648 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8668, lr: 0.0000, train/prefecth_time: 0.0044 
[2023-08-25 01:56:03,036 INFO] 59904 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6098, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 01:59:00,651 INFO] 60160 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6381, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 02:01:26,993 INFO] 60416 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3008, lr: 0.0000, train/prefecth_time: 0.0051 
[2023-08-25 02:03:53,577 INFO] 60672 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6131, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 02:06:53,464 INFO] 60928 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8179, lr: 0.0000, train/prefecth_time: 0.0049 
[2023-08-25 02:09:57,779 INFO] 61184 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6955, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 02:12:57,559 INFO] validating...
[2023-08-25 02:13:26,499 INFO] confusion matrix:
[[0.9228 0.028  0.0412 0.008 ]
 [0.0188 0.9772 0.0028 0.0012]
 [0.1788 0.0224 0.7056 0.0932]
 [0.1964 0.0188 0.1116 0.6732]]
[2023-08-25 02:13:28,803 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 02:13:28,804 INFO] 61440 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6415, eval/loss: 5.0696, eval/top-1-acc: 0.8197, eval/balanced_acc: 0.8197, eval/precision: 0.8305, eval/recall: 0.8197, eval/F1: 0.8170, lr: 0.0000, train/prefecth_time: 0.0025 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 02:15:12,102 INFO] 61696 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2976, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-25 02:18:24,889 INFO] 61952 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6111, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 02:21:23,261 INFO] 62208 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6759, lr: 0.0000, train/prefecth_time: 0.0043 
[2023-08-25 02:24:20,455 INFO] 62464 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6305, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 02:27:15,777 INFO] 62720 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2574, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 02:29:12,716 INFO] 62976 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7536, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-25 02:32:08,415 INFO] 63232 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6084, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 02:35:08,904 INFO] validating...
[2023-08-25 02:35:37,713 INFO] confusion matrix:
[[0.9176 0.0284 0.046  0.008 ]
 [0.0184 0.9776 0.0028 0.0012]
 [0.1668 0.02   0.7292 0.084 ]
 [0.1904 0.0184 0.128  0.6632]]
[2023-08-25 02:35:40,063 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 02:35:40,064 INFO] 63488 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7568, eval/loss: 5.1017, eval/top-1-acc: 0.8219, eval/balanced_acc: 0.8219, eval/precision: 0.8318, eval/recall: 0.8219, eval/F1: 0.8192, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 02:38:55,602 INFO] 63744 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6173, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 02:41:37,706 INFO] 64000 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3057, lr: 0.0000, train/prefecth_time: 0.0040 
[2023-08-25 02:43:41,794 INFO] 64256 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7547, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 02:46:38,917 INFO] 64512 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6321, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 02:49:36,054 INFO] 64768 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6475, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 02:52:33,602 INFO] 65024 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5154, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 02:54:38,187 INFO] 65280 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2518, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-25 02:57:19,776 INFO] validating...
[2023-08-25 02:57:49,153 INFO] confusion matrix:
[[0.9152 0.0276 0.0492 0.008 ]
 [0.0184 0.9776 0.0028 0.0012]
 [0.1584 0.0184 0.7512 0.072 ]
 [0.192  0.0184 0.1532 0.6364]]
[2023-08-25 02:57:51,654 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 02:57:51,655 INFO] 65536 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6174, eval/loss: 5.1975, eval/top-1-acc: 0.8201, eval/balanced_acc: 0.8201, eval/precision: 0.8308, eval/recall: 0.8201, eval/F1: 0.8170, lr: 0.0000, train/prefecth_time: 0.0025 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 03:01:00,052 INFO] 65792 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7804, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-25 03:04:25,856 INFO] 66048 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5949, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 03:07:17,510 INFO] 66304 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2896, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-25 03:08:56,969 INFO] 66560 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3385, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-25 03:11:40,408 INFO] 66816 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6971, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 03:14:42,466 INFO] 67072 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.6630, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 03:17:43,251 INFO] 67328 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6599, lr: 0.0000, train/prefecth_time: 0.0026 
[2023-08-25 03:20:42,488 INFO] validating...
[2023-08-25 03:21:02,016 INFO] confusion matrix:
[[0.9148 0.0264 0.0516 0.0072]
 [0.0192 0.9768 0.0028 0.0012]
 [0.1584 0.0152 0.7584 0.068 ]
 [0.1956 0.0172 0.1664 0.6208]]
[2023-08-25 03:21:04,926 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 03:21:04,927 INFO] 67584 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6067, eval/loss: 5.3015, eval/top-1-acc: 0.8177, eval/balanced_acc: 0.8177, eval/precision: 0.8296, eval/recall: 0.8177, eval/F1: 0.8143, lr: 0.0000, train/prefecth_time: 0.0046 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 03:23:31,210 INFO] 67840 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8356, lr: 0.0000, train/prefecth_time: 0.0035 
[2023-08-25 03:26:52,330 INFO] 68096 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6676, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 03:29:54,476 INFO] 68352 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6437, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 03:32:56,373 INFO] 68608 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6348, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 03:35:20,392 INFO] 68864 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3145, lr: 0.0000, train/prefecth_time: 0.0026 
[2023-08-25 03:37:30,357 INFO] 69120 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6184, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-25 03:40:30,477 INFO] 69376 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6349, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 03:43:26,556 INFO] validating...
[2023-08-25 03:43:55,571 INFO] confusion matrix:
[[0.9144 0.0264 0.0532 0.006 ]
 [0.02   0.9764 0.0024 0.0012]
 [0.1628 0.0156 0.7548 0.0668]
 [0.2148 0.0176 0.1676 0.6   ]]
[2023-08-25 03:43:58,115 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 03:43:58,116 INFO] 69632 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5119, eval/loss: 5.4226, eval/top-1-acc: 0.8114, eval/balanced_acc: 0.8114, eval/precision: 0.8254, eval/recall: 0.8114, eval/F1: 0.8075, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 03:47:14,565 INFO] 69888 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6299, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 03:49:24,898 INFO] 70144 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3548, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-25 03:52:16,392 INFO] 70400 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6433, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 03:55:16,293 INFO] 70656 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5642, lr: 0.0000, train/prefecth_time: 0.0067 
[2023-08-25 03:58:23,015 INFO] 70912 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7354, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 04:01:14,266 INFO] 71168 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3044, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-25 04:02:55,816 INFO] 71424 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2589, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 04:05:49,991 INFO] validating...
[2023-08-25 04:06:19,913 INFO] confusion matrix:
[[0.9196 0.0272 0.048  0.0052]
 [0.0196 0.9768 0.0024 0.0012]
 [0.1716 0.0168 0.7436 0.068 ]
 [0.2324 0.0176 0.1572 0.5928]]
[2023-08-25 04:06:22,467 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 04:06:22,468 INFO] 71680 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6186, eval/loss: 5.5109, eval/top-1-acc: 0.8082, eval/balanced_acc: 0.8082, eval/precision: 0.8239, eval/recall: 0.8082, eval/F1: 0.8042, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 04:09:45,319 INFO] 71936 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8225, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-25 04:12:56,974 INFO] 72192 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7016, lr: 0.0000, train/prefecth_time: 0.0042 
[2023-08-25 04:15:35,622 INFO] 72448 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2662, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 04:17:43,619 INFO] 72704 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6027, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 04:20:42,309 INFO] 72960 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5753, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 04:23:46,209 INFO] 73216 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6646, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 04:26:46,915 INFO] 73472 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7853, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-25 04:29:04,429 INFO] validating...
[2023-08-25 04:29:21,870 INFO] confusion matrix:
[[0.9156 0.0264 0.0512 0.0068]
 [0.0204 0.976  0.0024 0.0012]
 [0.1656 0.0132 0.75   0.0712]
 [0.2188 0.0168 0.1568 0.6076]]
[2023-08-25 04:29:24,287 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 04:29:24,288 INFO] 73728 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2734, eval/loss: 5.2627, eval/top-1-acc: 0.8123, eval/balanced_acc: 0.8123, eval/precision: 0.8261, eval/recall: 0.8123, eval/F1: 0.8088, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 04:32:23,873 INFO] 73984 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8189, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-25 04:35:29,240 INFO] 74240 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6047, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 04:38:29,406 INFO] 74496 iteration USE_EMA: True, train/sup_loss: 2.2231, train/unsup_loss: 0.0000, train/total_loss: 2.2231, train/util_ratio: 1.0000, train/run_time: 0.8009, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 04:41:26,806 INFO] 74752 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6268, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 04:43:20,068 INFO] 75008 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7102, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 04:46:16,729 INFO] 75264 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6976, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 04:49:16,995 INFO] 75520 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7015, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 04:52:20,702 INFO] validating...
[2023-08-25 04:52:51,009 INFO] confusion matrix:
[[0.9172 0.026  0.0508 0.006 ]
 [0.0204 0.9756 0.0028 0.0012]
 [0.1728 0.012  0.7476 0.0676]
 [0.2296 0.0152 0.1708 0.5844]]
[2023-08-25 04:52:53,683 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 04:52:53,684 INFO] 75776 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5952, eval/loss: 5.3199, eval/top-1-acc: 0.8062, eval/balanced_acc: 0.8062, eval/precision: 0.8221, eval/recall: 0.8062, eval/F1: 0.8021, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 04:55:59,574 INFO] 76032 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2878, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-25 04:57:51,967 INFO] 76288 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6351, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-25 05:00:49,581 INFO] 76544 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6769, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 05:03:45,778 INFO] 76800 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7110, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 05:06:46,758 INFO] 77056 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6138, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 05:09:08,270 INFO] 77312 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2762, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-25 05:11:34,526 INFO] 77568 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6487, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 05:14:30,462 INFO] validating...
[2023-08-25 05:14:59,155 INFO] confusion matrix:
[[0.9168 0.026  0.0508 0.0064]
 [0.0188 0.9772 0.0028 0.0012]
 [0.1756 0.0116 0.7412 0.0716]
 [0.2308 0.0156 0.1716 0.582 ]]
[2023-08-25 05:15:01,533 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 05:15:01,534 INFO] 77824 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6201, eval/loss: 5.3354, eval/top-1-acc: 0.8043, eval/balanced_acc: 0.8043, eval/precision: 0.8197, eval/recall: 0.8043, eval/F1: 0.8000, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 05:18:28,286 INFO] 78080 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8079, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-25 05:21:36,564 INFO] 78336 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0014, train/total_loss: 0.0014, train/util_ratio: 1.0000, train/run_time: 0.6496, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 05:23:26,267 INFO] 78592 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2871, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-25 05:26:06,676 INFO] 78848 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6346, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 05:29:06,925 INFO] 79104 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6016, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 05:32:04,147 INFO] 79360 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 0.8750, train/run_time: 0.5581, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 05:35:03,527 INFO] 79616 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6094, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 05:36:51,944 INFO] validating...
[2023-08-25 05:37:20,946 INFO] confusion matrix:
[[0.9184 0.0264 0.0472 0.008 ]
 [0.0192 0.9776 0.002  0.0012]
 [0.1816 0.0132 0.7176 0.0876]
 [0.2196 0.0148 0.1496 0.616 ]]
[2023-08-25 05:37:23,331 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 05:37:23,331 INFO] 79872 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5729, eval/loss: 5.3412, eval/top-1-acc: 0.8074, eval/balanced_acc: 0.8074, eval/precision: 0.8201, eval/recall: 0.8074, eval/F1: 0.8039, lr: 0.0000, train/prefecth_time: 0.0046 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 05:40:47,864 INFO] 80128 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6492, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 05:43:56,025 INFO] 80384 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6898, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 05:46:57,340 INFO] 80640 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6016, lr: 0.0000, train/prefecth_time: 0.0129 
[2023-08-25 05:49:29,956 INFO] 80896 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3454, lr: 0.0000, train/prefecth_time: 0.0027 
[2023-08-25 05:51:26,805 INFO] 81152 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6065, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 05:54:23,149 INFO] 81408 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6516, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 05:57:18,620 INFO] 81664 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6300, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 06:00:16,417 INFO] validating...
[2023-08-25 06:00:45,768 INFO] confusion matrix:
[[0.9136 0.0268 0.0504 0.0092]
 [0.018  0.9784 0.002  0.0016]
 [0.1728 0.0156 0.7148 0.0968]
 [0.2032 0.0164 0.1368 0.6436]]
[2023-08-25 06:00:48,203 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 06:00:48,204 INFO] 81920 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6247, eval/loss: 5.3835, eval/top-1-acc: 0.8126, eval/balanced_acc: 0.8126, eval/precision: 0.8224, eval/recall: 0.8126, eval/F1: 0.8096, lr: 0.0000, train/prefecth_time: 0.0045 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 06:03:04,959 INFO] 82176 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2751, lr: 0.0000, train/prefecth_time: 0.0040 
[2023-08-25 06:05:55,901 INFO] 82432 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7301, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 06:08:47,660 INFO] 82688 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6623, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 06:11:49,942 INFO] 82944 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6798, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 06:14:47,376 INFO] 83200 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.4326, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 06:16:33,767 INFO] 83456 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3288, lr: 0.0000, train/prefecth_time: 0.0034 
[2023-08-25 06:19:09,534 INFO] 83712 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7483, lr: 0.0000, train/prefecth_time: 0.0038 
[2023-08-25 06:22:01,532 INFO] validating...
[2023-08-25 06:22:30,132 INFO] confusion matrix:
[[0.9108 0.0276 0.0492 0.0124]
 [0.016  0.9808 0.0016 0.0016]
 [0.1668 0.0168 0.6984 0.118 ]
 [0.1776 0.0168 0.118  0.6876]]
[2023-08-25 06:22:32,576 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 06:22:32,577 INFO] 83968 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7811, eval/loss: 5.3659, eval/top-1-acc: 0.8194, eval/balanced_acc: 0.8194, eval/precision: 0.8255, eval/recall: 0.8194, eval/F1: 0.8166, lr: 0.0000, train/prefecth_time: 0.0045 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 06:25:48,899 INFO] 84224 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8150, lr: 0.0000, train/prefecth_time: 0.0035 
[2023-08-25 06:28:55,385 INFO] 84480 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5471, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 06:30:40,993 INFO] 84736 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7116, lr: 0.0000, train/prefecth_time: 0.0040 
[2023-08-25 06:33:37,477 INFO] 84992 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5779, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 06:35:26,953 INFO] 85248 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2755, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-25 06:37:16,163 INFO] 85504 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3219, lr: 0.0000, train/prefecth_time: 0.0029 
[2023-08-25 06:39:50,683 INFO] 85760 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7313, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-25 06:42:47,739 INFO] validating...
[2023-08-25 06:43:17,529 INFO] confusion matrix:
[[0.9096 0.028  0.0472 0.0152]
 [0.014  0.9824 0.0016 0.002 ]
 [0.1688 0.0172 0.6836 0.1304]
 [0.1592 0.016  0.1044 0.7204]]
[2023-08-25 06:43:19,884 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 06:43:19,885 INFO] 86016 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6946, eval/loss: 5.3920, eval/top-1-acc: 0.8240, eval/balanced_acc: 0.8240, eval/precision: 0.8287, eval/recall: 0.8240, eval/F1: 0.8213, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 06:46:29,214 INFO] 86272 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6854, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 06:49:46,937 INFO] 86528 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2751, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 06:51:45,921 INFO] 86784 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7077, lr: 0.0000, train/prefecth_time: 0.0028 
[2023-08-25 06:54:44,548 INFO] 87040 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6635, lr: 0.0000, train/prefecth_time: 0.0074 
[2023-08-25 06:57:45,531 INFO] 87296 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.7164, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 07:00:43,533 INFO] 87552 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5740, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 07:03:00,009 INFO] 87808 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3048, lr: 0.0000, train/prefecth_time: 0.0040 
[2023-08-25 07:05:00,074 INFO] validating...
[2023-08-25 07:05:21,736 INFO] confusion matrix:
[[0.9104 0.0288 0.044  0.0168]
 [0.0124 0.9836 0.0016 0.0024]
 [0.1708 0.0176 0.6716 0.14  ]
 [0.1436 0.0156 0.0916 0.7492]]
[2023-08-25 07:05:29,156 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 07:05:29,157 INFO] 88064 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3120, eval/loss: 5.4457, eval/top-1-acc: 0.8287, eval/balanced_acc: 0.8287, eval/precision: 0.8329, eval/recall: 0.8287, eval/F1: 0.8258, lr: 0.0000, train/prefecth_time: 0.0037 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 07:07:42,450 INFO] 88320 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8367, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 07:11:42,756 INFO] 88576 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8776, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 07:14:39,598 INFO] 88832 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2868, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-25 07:17:52,452 INFO] 89088 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8286, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 07:21:55,289 INFO] 89344 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8223, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 07:23:44,606 INFO] 89600 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0007, train/total_loss: 0.0007, train/util_ratio: 1.0000, train/run_time: 0.3135, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-25 07:27:27,176 INFO] 89856 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8738, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 07:31:12,155 INFO] validating...
[2023-08-25 07:31:29,137 INFO] confusion matrix:
[[0.9124 0.0292 0.042  0.0164]
 [0.0128 0.9832 0.0016 0.0024]
 [0.1744 0.0168 0.668  0.1408]
 [0.1412 0.0164 0.0872 0.7552]]
[2023-08-25 07:31:31,778 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 07:31:31,778 INFO] 90112 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2813, eval/loss: 5.4975, eval/top-1-acc: 0.8297, eval/balanced_acc: 0.8297, eval/precision: 0.8344, eval/recall: 0.8297, eval/F1: 0.8268, lr: 0.0000, train/prefecth_time: 0.0032 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 07:34:23,121 INFO] 90368 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9064, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 07:38:25,698 INFO] 90624 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8638, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 07:40:51,344 INFO] 90880 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3220, lr: 0.0000, train/prefecth_time: 0.0037 
[2023-08-25 07:43:03,347 INFO] 91136 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 1.0391, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 07:47:03,686 INFO] 91392 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8853, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-25 07:49:47,183 INFO] 91648 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3078, lr: 0.0000, train/prefecth_time: 0.0032 
[2023-08-25 07:53:07,169 INFO] 91904 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9168, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 07:57:04,520 INFO] validating...
[2023-08-25 07:57:26,732 INFO] confusion matrix:
[[0.9132 0.0292 0.0412 0.0164]
 [0.012  0.984  0.0016 0.0024]
 [0.1812 0.0172 0.6596 0.142 ]
 [0.14   0.0172 0.0844 0.7584]]
[2023-08-25 07:57:29,552 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 07:57:29,553 INFO] 92160 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8926, eval/loss: 5.6251, eval/top-1-acc: 0.8288, eval/balanced_acc: 0.8288, eval/precision: 0.8338, eval/recall: 0.8288, eval/F1: 0.8257, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 07:59:29,549 INFO] 92416 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 1.2300, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 08:03:32,991 INFO] 92672 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9184, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 08:06:48,434 INFO] 92928 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3168, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 08:09:42,858 INFO] 93184 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8583, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 08:13:42,653 INFO] 93440 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8350, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 08:15:53,158 INFO] 93696 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3127, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 08:19:19,045 INFO] 93952 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8754, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 08:23:20,699 INFO] validating...
[2023-08-25 08:23:38,562 INFO] confusion matrix:
[[0.9144 0.03   0.0396 0.016 ]
 [0.0108 0.9856 0.0016 0.002 ]
 [0.184  0.0188 0.658  0.1392]
 [0.1396 0.0192 0.0872 0.754 ]]
[2023-08-25 08:23:41,175 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 08:23:41,176 INFO] 94208 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9016, eval/loss: 5.7945, eval/top-1-acc: 0.8280, eval/balanced_acc: 0.8280, eval/precision: 0.8330, eval/recall: 0.8280, eval/F1: 0.8247, lr: 0.0000, train/prefecth_time: 0.0040 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 08:26:32,390 INFO] 94464 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8493, lr: 0.0000, train/prefecth_time: 0.0026 
[2023-08-25 08:30:32,690 INFO] 94720 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9943, lr: 0.0000, train/prefecth_time: 0.0030 
[2023-08-25 08:32:55,531 INFO] 94976 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2933, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 08:36:02,426 INFO] 95232 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9160, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 08:39:57,037 INFO] 95488 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8953, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 08:42:04,575 INFO] 95744 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9144, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 08:46:02,488 INFO] 96000 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8890, lr: 0.0000, train/prefecth_time: 0.0036 
[2023-08-25 08:49:29,216 INFO] validating...
[2023-08-25 08:49:50,474 INFO] confusion matrix:
[[0.9128 0.0316 0.0396 0.016 ]
 [0.0092 0.9872 0.0016 0.002 ]
 [0.1856 0.0232 0.6516 0.1396]
 [0.138  0.0256 0.0884 0.748 ]]
[2023-08-25 08:49:52,933 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 08:49:52,934 INFO] 96256 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8071, eval/loss: 6.0630, eval/top-1-acc: 0.8249, eval/balanced_acc: 0.8249, eval/precision: 0.8294, eval/recall: 0.8249, eval/F1: 0.8211, lr: 0.0000, train/prefecth_time: 0.0057 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 08:52:31,229 INFO] 96512 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9108, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 08:56:32,550 INFO] 96768 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8546, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 08:58:50,492 INFO] 97024 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2809, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 09:02:44,787 INFO] 97280 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8906, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:06:39,195 INFO] 97536 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2905, lr: 0.0000, train/prefecth_time: 0.0033 
[2023-08-25 09:08:19,636 INFO] 97792 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 1.3085, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:12:20,538 INFO] 98048 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9136, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 09:15:31,685 INFO] validating...
[2023-08-25 09:15:48,942 INFO] confusion matrix:
[[0.9128 0.0328 0.0392 0.0152]
 [0.0088 0.9876 0.0016 0.002 ]
 [0.186  0.0288 0.64   0.1452]
 [0.1376 0.0292 0.0832 0.75  ]]
[2023-08-25 09:15:51,308 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 09:15:51,309 INFO] 98304 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.2912, eval/loss: 6.3269, eval/top-1-acc: 0.8226, eval/balanced_acc: 0.8226, eval/precision: 0.8271, eval/recall: 0.8226, eval/F1: 0.8184, lr: 0.0000, train/prefecth_time: 0.0028 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 09:19:27,541 INFO] 98560 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9895, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:22:50,975 INFO] 98816 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8479, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 09:24:56,609 INFO] 99072 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9682, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:28:52,997 INFO] 99328 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 1.0077, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 09:31:12,982 INFO] 99584 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3155, lr: 0.0000, train/prefecth_time: 0.0046 
[2023-08-25 09:34:53,453 INFO] 99840 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8486, lr: 0.0000, train/prefecth_time: 0.0025 
[2023-08-25 09:38:18,398 INFO] 100096 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8816, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:40:27,185 INFO] validating...
[2023-08-25 09:41:06,537 INFO] confusion matrix:
[[0.91   0.0344 0.0392 0.0164]
 [0.0084 0.988  0.0016 0.002 ]
 [0.1852 0.0332 0.626  0.1556]
 [0.1344 0.0316 0.0768 0.7572]]
[2023-08-25 09:41:08,789 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 09:41:08,790 INFO] 100352 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8735, eval/loss: 6.5569, eval/top-1-acc: 0.8203, eval/balanced_acc: 0.8203, eval/precision: 0.8247, eval/recall: 0.8203, eval/F1: 0.8155, lr: 0.0000, train/prefecth_time: 0.0045 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 09:45:07,416 INFO] 100608 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.8481, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:47:16,411 INFO] 100864 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.9952, lr: 0.0000, train/prefecth_time: 0.0038 
[2023-08-25 09:50:51,188 INFO] 101120 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6041, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:53:38,776 INFO] 101376 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.5581, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 09:55:35,194 INFO] 101632 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.3478, lr: 0.0000, train/prefecth_time: 0.0031 
[2023-08-25 09:58:08,093 INFO] 101888 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6292, lr: 0.0000, train/prefecth_time: 0.0045 
[2023-08-25 10:00:54,280 INFO] 102144 iteration USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6128, lr: 0.0000, train/prefecth_time: 0.0024 
[2023-08-25 10:03:43,705 INFO] validating...
[2023-08-25 10:04:12,326 INFO] confusion matrix:
[[0.9096 0.0348 0.0392 0.0164]
 [0.0084 0.988  0.0016 0.002 ]
 [0.1824 0.0348 0.6176 0.1652]
 [0.13   0.0332 0.072  0.7648]]
[2023-08-25 10:04:14,665 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 10:04:14,666 INFO] 102400 iteration, USE_EMA: True, train/sup_loss: 0.0000, train/unsup_loss: 0.0000, train/total_loss: 0.0000, train/util_ratio: 1.0000, train/run_time: 0.6159, eval/loss: 6.7708, eval/top-1-acc: 0.8200, eval/balanced_acc: 0.8200, eval/precision: 0.8243, eval/recall: 0.8200, eval/F1: 0.8149, lr: 0.0000, train/prefecth_time: 0.0024 BEST_EVAL_ACC: 0.8741, at 26624 iters
[2023-08-25 10:04:17,797 INFO] model saved: /liuzicheng/jwy/saved_models/usb_nlp/pseudolabel_ag_news_40_0/latest_model.pth
[2023-08-25 10:04:26,133 INFO] Model loaded
[2023-08-25 10:04:48,999 INFO] confusion matrix:
[[0.90052632 0.03947368 0.03736842 0.02263158]
 [0.01       0.98736842 0.00105263 0.00157895]
 [0.16736842 0.03894737 0.61947368 0.17421053]
 [0.12789474 0.02947368 0.07526316 0.76736842]]
[2023-08-25 10:04:49,003 INFO] Model result - eval/best_acc : 0.8741
[2023-08-25 10:04:49,004 INFO] Model result - eval/best_it : 26623

