==========
Args:Namespace(dataset='market1501', batch_size=256, workers=4, height=256, width=128, num_instances=8, eps=0.6, eps_gap=0.02, k1=30, k2=6, arch='vit_small', pretrained_path='/home/ma-user/work/Projects/ReIDNet_Finetune/TransReID/log/bs_exp/ViT_Small_MSMT17/stem_64bs_lr0.0004_ep120_warm20_seed0/vit_small_120.pth', features=0, dropout=0, momentum=0.2, drop_path_rate=0.3, hw_ratio=2, self_norm=True, conv_stem=True, lr=0.00035, weight_decay=0.0005, optimizer='SGD', epochs=50, iters=200, step_size=20, seed=1, print_freq=20, eval_step=10, temp=0.05, data_dir='/home/ma-user/work/Projects/ReIDNet_Finetune/CContrast/data', logs_dir='log/msmt2market/vit_small_cion', pooling_type='gem', use_hard=True)
==========
==> Load unlabeled dataset
=> Market1501 loaded
Dataset statistics:
  ----------------------------------------
  subset   | # ids | # images | # cameras
  ----------------------------------------
  train    |   751 |    12936 |         6
  query    |   750 |     3368 |         6
  gallery  |   751 |    15913 |         6
  ----------------------------------------
Using convolution stem
using drop_out rate is : 0.0
using attn_drop_out rate is : 0.0
using drop_path rate is : 0.3
Load 172 / 177 layers.
optimizer: SGD
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.467)	Data 0.000 (0.029)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.14078712463379
Clustering criterion: eps: 0.600
==> Statistics for epoch 0: 456 clusters
Epoch: [0][20/200]	Time 0.265 (0.878)	Data 0.001 (0.096)	Loss 5.878 (5.800)
Epoch: [0][40/200]	Time 0.265 (0.603)	Data 0.000 (0.079)	Loss 3.698 (5.379)
Epoch: [0][60/200]	Time 0.272 (0.539)	Data 0.001 (0.098)	Loss 2.889 (4.795)
Epoch: [0][80/200]	Time 0.262 (0.487)	Data 0.000 (0.089)	Loss 2.963 (4.420)
Epoch: [0][100/200]	Time 0.270 (0.472)	Data 0.000 (0.099)	Loss 2.923 (4.155)
Epoch: [0][120/200]	Time 0.263 (0.449)	Data 0.000 (0.094)	Loss 2.906 (3.952)
Epoch: [0][140/200]	Time 0.267 (0.432)	Data 0.000 (0.089)	Loss 3.006 (3.783)
Epoch: [0][160/200]	Time 0.264 (0.429)	Data 0.001 (0.095)	Loss 2.504 (3.645)
Epoch: [0][180/200]	Time 0.264 (0.419)	Data 0.000 (0.092)	Loss 2.759 (3.524)
Epoch: [0][200/200]	Time 0.270 (0.419)	Data 0.001 (0.097)	Loss 2.199 (3.416)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.152)	Data 0.000 (0.028)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.27944779396057
==> Statistics for epoch 1: 527 clusters
Epoch: [1][20/200]	Time 0.263 (0.380)	Data 0.001 (0.109)	Loss 2.322 (0.879)
Epoch: [1][40/200]	Time 0.264 (0.363)	Data 0.001 (0.090)	Loss 2.525 (1.600)
Epoch: [1][60/200]	Time 0.267 (0.358)	Data 0.000 (0.086)	Loss 2.346 (1.873)
Epoch: [1][80/200]	Time 0.266 (0.355)	Data 0.000 (0.082)	Loss 2.313 (2.006)
Epoch: [1][100/200]	Time 0.266 (0.369)	Data 0.001 (0.096)	Loss 1.786 (2.056)
Epoch: [1][120/200]	Time 0.267 (0.365)	Data 0.001 (0.093)	Loss 2.203 (2.109)
Epoch: [1][140/200]	Time 0.266 (0.363)	Data 0.000 (0.090)	Loss 2.466 (2.134)
Epoch: [1][160/200]	Time 0.262 (0.360)	Data 0.000 (0.088)	Loss 2.171 (2.134)
Epoch: [1][180/200]	Time 0.266 (0.366)	Data 0.001 (0.093)	Loss 2.131 (2.125)
Epoch: [1][200/200]	Time 0.267 (0.365)	Data 0.001 (0.092)	Loss 2.261 (2.124)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.134)	Data 0.000 (0.032)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.93046236038208
==> Statistics for epoch 2: 553 clusters
Epoch: [2][20/200]	Time 0.263 (0.391)	Data 0.001 (0.115)	Loss 2.201 (0.611)
Epoch: [2][40/200]	Time 0.267 (0.363)	Data 0.001 (0.090)	Loss 1.964 (1.323)
Epoch: [2][60/200]	Time 0.266 (0.360)	Data 0.001 (0.086)	Loss 1.505 (1.562)
Epoch: [2][80/200]	Time 0.264 (0.357)	Data 0.000 (0.083)	Loss 1.973 (1.691)
Epoch: [2][100/200]	Time 0.266 (0.355)	Data 0.000 (0.081)	Loss 2.168 (1.759)
Epoch: [2][120/200]	Time 1.797 (0.366)	Data 1.518 (0.092)	Loss 2.652 (1.786)
Epoch: [2][140/200]	Time 0.271 (0.362)	Data 0.001 (0.089)	Loss 2.187 (1.808)
Epoch: [2][160/200]	Time 0.269 (0.360)	Data 0.001 (0.086)	Loss 1.922 (1.821)
Epoch: [2][180/200]	Time 0.267 (0.359)	Data 0.000 (0.084)	Loss 2.279 (1.842)
Epoch: [2][200/200]	Time 0.269 (0.357)	Data 0.000 (0.083)	Loss 1.722 (1.846)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.138)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.036505222320557
==> Statistics for epoch 3: 558 clusters
Epoch: [3][20/200]	Time 0.266 (0.373)	Data 0.001 (0.104)	Loss 2.046 (0.570)
Epoch: [3][40/200]	Time 0.267 (0.357)	Data 0.001 (0.085)	Loss 1.570 (1.211)
Epoch: [3][60/200]	Time 0.267 (0.351)	Data 0.001 (0.079)	Loss 1.579 (1.408)
Epoch: [3][80/200]	Time 0.297 (0.347)	Data 0.000 (0.075)	Loss 1.762 (1.530)
Epoch: [3][100/200]	Time 0.266 (0.345)	Data 0.000 (0.074)	Loss 1.454 (1.570)
Epoch: [3][120/200]	Time 1.708 (0.356)	Data 1.413 (0.085)	Loss 1.953 (1.624)
Epoch: [3][140/200]	Time 0.267 (0.354)	Data 0.001 (0.082)	Loss 1.725 (1.648)
Epoch: [3][160/200]	Time 0.272 (0.352)	Data 0.001 (0.080)	Loss 1.352 (1.654)
Epoch: [3][180/200]	Time 0.267 (0.351)	Data 0.001 (0.079)	Loss 1.383 (1.656)
Epoch: [3][200/200]	Time 0.263 (0.349)	Data 0.000 (0.078)	Loss 2.084 (1.666)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.138)	Data 0.000 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.04752826690674
==> Statistics for epoch 4: 564 clusters
Epoch: [4][20/200]	Time 0.274 (0.367)	Data 0.001 (0.097)	Loss 1.912 (0.488)
Epoch: [4][40/200]	Time 0.265 (0.354)	Data 0.001 (0.080)	Loss 1.691 (1.087)
Epoch: [4][60/200]	Time 0.269 (0.350)	Data 0.001 (0.076)	Loss 1.729 (1.317)
Epoch: [4][80/200]	Time 0.266 (0.348)	Data 0.000 (0.074)	Loss 1.779 (1.413)
Epoch: [4][100/200]	Time 0.268 (0.344)	Data 0.000 (0.072)	Loss 1.309 (1.459)
Epoch: [4][120/200]	Time 1.626 (0.354)	Data 1.284 (0.081)	Loss 1.381 (1.498)
Epoch: [4][140/200]	Time 0.265 (0.352)	Data 0.001 (0.079)	Loss 1.878 (1.523)
Epoch: [4][160/200]	Time 0.267 (0.350)	Data 0.001 (0.077)	Loss 1.199 (1.540)
Epoch: [4][180/200]	Time 0.270 (0.348)	Data 0.001 (0.075)	Loss 1.640 (1.561)
Epoch: [4][200/200]	Time 0.268 (0.347)	Data 0.000 (0.074)	Loss 1.430 (1.564)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.126 (0.132)	Data 0.030 (0.032)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.286227226257324
==> Statistics for epoch 5: 573 clusters
Epoch: [5][20/200]	Time 0.266 (0.367)	Data 0.001 (0.095)	Loss 1.115 (0.439)
Epoch: [5][40/200]	Time 0.267 (0.352)	Data 0.001 (0.080)	Loss 1.590 (0.959)
Epoch: [5][60/200]	Time 0.267 (0.351)	Data 0.001 (0.078)	Loss 1.714 (1.168)
Epoch: [5][80/200]	Time 0.267 (0.349)	Data 0.000 (0.077)	Loss 1.710 (1.300)
Epoch: [5][100/200]	Time 0.264 (0.349)	Data 0.000 (0.076)	Loss 1.578 (1.347)
Epoch: [5][120/200]	Time 1.686 (0.359)	Data 1.386 (0.086)	Loss 1.614 (1.382)
Epoch: [5][140/200]	Time 0.269 (0.357)	Data 0.001 (0.084)	Loss 1.607 (1.412)
Epoch: [5][160/200]	Time 0.264 (0.354)	Data 0.001 (0.081)	Loss 2.195 (1.436)
Epoch: [5][180/200]	Time 0.265 (0.352)	Data 0.000 (0.080)	Loss 1.115 (1.449)
Epoch: [5][200/200]	Time 0.269 (0.351)	Data 0.000 (0.079)	Loss 1.672 (1.455)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.186 (0.132)	Data 0.000 (0.032)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.190950632095337
==> Statistics for epoch 6: 578 clusters
Epoch: [6][20/200]	Time 0.266 (0.376)	Data 0.001 (0.103)	Loss 1.474 (0.386)
Epoch: [6][40/200]	Time 0.267 (0.353)	Data 0.000 (0.082)	Loss 1.480 (0.938)
Epoch: [6][60/200]	Time 0.266 (0.350)	Data 0.001 (0.078)	Loss 1.653 (1.138)
Epoch: [6][80/200]	Time 0.268 (0.348)	Data 0.001 (0.076)	Loss 1.665 (1.229)
Epoch: [6][100/200]	Time 0.270 (0.348)	Data 0.000 (0.075)	Loss 1.327 (1.289)
Epoch: [6][120/200]	Time 0.267 (0.347)	Data 0.000 (0.074)	Loss 1.601 (1.344)
Epoch: [6][140/200]	Time 0.263 (0.345)	Data 0.000 (0.073)	Loss 1.561 (1.377)
Epoch: [6][160/200]	Time 0.268 (0.345)	Data 0.000 (0.073)	Loss 1.735 (1.379)
Epoch: [6][180/200]	Time 0.263 (0.345)	Data 0.000 (0.073)	Loss 1.215 (1.384)
Epoch: [6][200/200]	Time 0.292 (0.351)	Data 0.001 (0.079)	Loss 1.421 (1.389)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.131)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.19779324531555
==> Statistics for epoch 7: 583 clusters
Epoch: [7][20/200]	Time 0.266 (0.381)	Data 0.000 (0.110)	Loss 1.583 (0.353)
Epoch: [7][40/200]	Time 0.267 (0.363)	Data 0.001 (0.090)	Loss 1.180 (0.844)
Epoch: [7][60/200]	Time 0.414 (0.359)	Data 0.001 (0.085)	Loss 1.901 (1.076)
Epoch: [7][80/200]	Time 0.273 (0.356)	Data 0.001 (0.083)	Loss 1.445 (1.160)
Epoch: [7][100/200]	Time 0.265 (0.354)	Data 0.000 (0.080)	Loss 1.141 (1.197)
Epoch: [7][120/200]	Time 0.268 (0.353)	Data 0.000 (0.079)	Loss 1.444 (1.242)
Epoch: [7][140/200]	Time 0.269 (0.351)	Data 0.000 (0.077)	Loss 1.420 (1.272)
Epoch: [7][160/200]	Time 0.267 (0.350)	Data 0.000 (0.077)	Loss 1.369 (1.296)
Epoch: [7][180/200]	Time 0.270 (0.351)	Data 0.000 (0.077)	Loss 1.001 (1.306)
Epoch: [7][200/200]	Time 0.270 (0.357)	Data 0.001 (0.083)	Loss 1.551 (1.313)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.134)	Data 0.000 (0.034)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.424811840057373
==> Statistics for epoch 8: 576 clusters
Epoch: [8][20/200]	Time 0.282 (0.374)	Data 0.001 (0.104)	Loss 1.374 (0.342)
Epoch: [8][40/200]	Time 0.282 (0.359)	Data 0.001 (0.086)	Loss 1.319 (0.838)
Epoch: [8][60/200]	Time 0.266 (0.352)	Data 0.001 (0.080)	Loss 1.332 (1.005)
Epoch: [8][80/200]	Time 0.268 (0.348)	Data 0.001 (0.077)	Loss 1.350 (1.088)
Epoch: [8][100/200]	Time 0.268 (0.346)	Data 0.000 (0.075)	Loss 1.133 (1.162)
Epoch: [8][120/200]	Time 0.267 (0.345)	Data 0.000 (0.073)	Loss 1.102 (1.207)
Epoch: [8][140/200]	Time 0.266 (0.344)	Data 0.000 (0.072)	Loss 1.012 (1.236)
Epoch: [8][160/200]	Time 0.269 (0.344)	Data 0.000 (0.072)	Loss 1.457 (1.261)
Epoch: [8][180/200]	Time 0.268 (0.344)	Data 0.000 (0.072)	Loss 1.471 (1.282)
Epoch: [8][200/200]	Time 0.370 (0.351)	Data 0.000 (0.078)	Loss 0.968 (1.281)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.135)	Data 0.000 (0.036)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.156575202941895
==> Statistics for epoch 9: 589 clusters
Epoch: [9][20/200]	Time 0.269 (0.372)	Data 0.001 (0.101)	Loss 1.396 (0.313)
Epoch: [9][40/200]	Time 0.272 (0.357)	Data 0.001 (0.086)	Loss 1.452 (0.794)
Epoch: [9][60/200]	Time 0.267 (0.351)	Data 0.001 (0.079)	Loss 1.389 (0.987)
Epoch: [9][80/200]	Time 0.268 (0.347)	Data 0.001 (0.076)	Loss 1.627 (1.076)
Epoch: [9][100/200]	Time 0.267 (0.346)	Data 0.001 (0.075)	Loss 1.054 (1.110)
Epoch: [9][120/200]	Time 0.266 (0.345)	Data 0.000 (0.074)	Loss 1.193 (1.150)
Epoch: [9][140/200]	Time 0.265 (0.345)	Data 0.000 (0.073)	Loss 1.081 (1.160)
Epoch: [9][160/200]	Time 0.263 (0.344)	Data 0.000 (0.073)	Loss 1.072 (1.177)
Epoch: [9][180/200]	Time 0.265 (0.344)	Data 0.000 (0.072)	Loss 1.545 (1.190)
Epoch: [9][200/200]	Time 0.272 (0.350)	Data 0.001 (0.079)	Loss 1.151 (1.194)
Extract Features: [50/76]	Time 0.098 (0.132)	Data 0.000 (0.033)	
Mean AP: 89.2%

 * Finished epoch   9  model mAP: 89.2%  best: 89.2% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.138)	Data 0.000 (0.034)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.907575607299805
==> Statistics for epoch 10: 583 clusters
Epoch: [10][20/200]	Time 0.265 (0.387)	Data 0.000 (0.093)	Loss 1.206 (0.290)
Epoch: [10][40/200]	Time 0.266 (0.363)	Data 0.001 (0.083)	Loss 1.204 (0.750)
Epoch: [10][60/200]	Time 0.275 (0.357)	Data 0.001 (0.079)	Loss 1.159 (0.911)
Epoch: [10][80/200]	Time 0.268 (0.353)	Data 0.001 (0.075)	Loss 1.081 (0.994)
Epoch: [10][100/200]	Time 0.282 (0.352)	Data 0.000 (0.075)	Loss 1.449 (1.064)
Epoch: [10][120/200]	Time 0.268 (0.350)	Data 0.000 (0.073)	Loss 1.313 (1.087)
Epoch: [10][140/200]	Time 0.263 (0.348)	Data 0.000 (0.072)	Loss 1.113 (1.105)
Epoch: [10][160/200]	Time 0.265 (0.347)	Data 0.000 (0.071)	Loss 1.750 (1.119)
Epoch: [10][180/200]	Time 0.265 (0.346)	Data 0.000 (0.071)	Loss 1.287 (1.141)
Epoch: [10][200/200]	Time 0.268 (0.352)	Data 0.000 (0.077)	Loss 1.424 (1.155)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.139)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.703699588775635
==> Statistics for epoch 11: 586 clusters
Epoch: [11][20/200]	Time 0.271 (0.386)	Data 0.000 (0.109)	Loss 1.294 (0.273)
Epoch: [11][40/200]	Time 0.266 (0.365)	Data 0.001 (0.091)	Loss 1.659 (0.705)
Epoch: [11][60/200]	Time 0.266 (0.360)	Data 0.001 (0.084)	Loss 1.167 (0.869)
Epoch: [11][80/200]	Time 0.268 (0.355)	Data 0.001 (0.081)	Loss 1.464 (0.959)
Epoch: [11][100/200]	Time 0.265 (0.353)	Data 0.001 (0.079)	Loss 1.095 (0.995)
Epoch: [11][120/200]	Time 0.265 (0.352)	Data 0.000 (0.078)	Loss 0.770 (1.020)
Epoch: [11][140/200]	Time 0.270 (0.352)	Data 0.000 (0.077)	Loss 1.135 (1.046)
Epoch: [11][160/200]	Time 0.268 (0.351)	Data 0.000 (0.077)	Loss 1.051 (1.069)
Epoch: [11][180/200]	Time 0.264 (0.351)	Data 0.000 (0.077)	Loss 1.302 (1.080)
Epoch: [11][200/200]	Time 0.273 (0.358)	Data 0.001 (0.084)	Loss 1.097 (1.091)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.135)	Data 0.000 (0.036)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.34424877166748
==> Statistics for epoch 12: 585 clusters
Epoch: [12][20/200]	Time 0.268 (0.377)	Data 0.001 (0.101)	Loss 1.198 (0.262)
Epoch: [12][40/200]	Time 0.272 (0.359)	Data 0.001 (0.083)	Loss 1.109 (0.673)
Epoch: [12][60/200]	Time 0.265 (0.354)	Data 0.001 (0.079)	Loss 0.887 (0.824)
Epoch: [12][80/200]	Time 0.268 (0.352)	Data 0.001 (0.078)	Loss 1.080 (0.899)
Epoch: [12][100/200]	Time 0.267 (0.351)	Data 0.001 (0.077)	Loss 1.037 (0.951)
Epoch: [12][120/200]	Time 0.267 (0.351)	Data 0.000 (0.077)	Loss 0.919 (0.970)
Epoch: [12][140/200]	Time 0.267 (0.350)	Data 0.000 (0.076)	Loss 1.191 (0.987)
Epoch: [12][160/200]	Time 0.267 (0.349)	Data 0.000 (0.076)	Loss 1.266 (1.011)
Epoch: [12][180/200]	Time 0.268 (0.349)	Data 0.000 (0.076)	Loss 1.156 (1.033)
Epoch: [12][200/200]	Time 0.269 (0.356)	Data 0.001 (0.082)	Loss 1.304 (1.042)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.139)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.95325756072998
==> Statistics for epoch 13: 587 clusters
Epoch: [13][20/200]	Time 0.270 (0.386)	Data 0.001 (0.109)	Loss 0.818 (0.253)
Epoch: [13][40/200]	Time 0.272 (0.364)	Data 0.001 (0.090)	Loss 1.144 (0.643)
Epoch: [13][60/200]	Time 0.268 (0.355)	Data 0.001 (0.082)	Loss 1.175 (0.808)
Epoch: [13][80/200]	Time 0.266 (0.352)	Data 0.001 (0.080)	Loss 1.386 (0.898)
Epoch: [13][100/200]	Time 0.273 (0.350)	Data 0.001 (0.077)	Loss 0.980 (0.943)
Epoch: [13][120/200]	Time 0.268 (0.350)	Data 0.000 (0.077)	Loss 0.783 (0.966)
Epoch: [13][140/200]	Time 0.263 (0.349)	Data 0.000 (0.076)	Loss 1.149 (0.991)
Epoch: [13][160/200]	Time 0.267 (0.347)	Data 0.000 (0.075)	Loss 1.241 (1.009)
Epoch: [13][180/200]	Time 0.267 (0.348)	Data 0.000 (0.075)	Loss 1.170 (1.014)
Epoch: [13][200/200]	Time 0.280 (0.355)	Data 0.001 (0.082)	Loss 0.911 (1.023)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.144)	Data 0.000 (0.041)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.107765913009644
==> Statistics for epoch 14: 585 clusters
Epoch: [14][20/200]	Time 0.277 (0.390)	Data 0.002 (0.110)	Loss 1.076 (0.265)
Epoch: [14][40/200]	Time 0.277 (0.368)	Data 0.001 (0.090)	Loss 1.152 (0.659)
Epoch: [14][60/200]	Time 0.270 (0.360)	Data 0.001 (0.085)	Loss 0.953 (0.804)
Epoch: [14][80/200]	Time 0.266 (0.356)	Data 0.001 (0.082)	Loss 0.841 (0.881)
Epoch: [14][100/200]	Time 0.269 (0.355)	Data 0.001 (0.079)	Loss 1.021 (0.935)
Epoch: [14][120/200]	Time 0.269 (0.354)	Data 0.000 (0.079)	Loss 0.915 (0.971)
Epoch: [14][140/200]	Time 0.267 (0.352)	Data 0.000 (0.077)	Loss 1.294 (0.982)
Epoch: [14][160/200]	Time 0.265 (0.352)	Data 0.000 (0.077)	Loss 1.159 (1.001)
Epoch: [14][180/200]	Time 0.360 (0.352)	Data 0.000 (0.077)	Loss 1.175 (1.019)
Epoch: [14][200/200]	Time 0.268 (0.358)	Data 0.001 (0.084)	Loss 1.289 (1.033)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.140)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.094172477722168
==> Statistics for epoch 15: 586 clusters
Epoch: [15][20/200]	Time 0.273 (0.378)	Data 0.002 (0.105)	Loss 1.084 (0.277)
Epoch: [15][40/200]	Time 0.267 (0.358)	Data 0.001 (0.087)	Loss 0.981 (0.666)
Epoch: [15][60/200]	Time 0.268 (0.354)	Data 0.001 (0.082)	Loss 0.771 (0.815)
Epoch: [15][80/200]	Time 0.268 (0.352)	Data 0.001 (0.079)	Loss 0.977 (0.874)
Epoch: [15][100/200]	Time 0.265 (0.351)	Data 0.001 (0.079)	Loss 1.008 (0.885)
Epoch: [15][120/200]	Time 0.265 (0.350)	Data 0.000 (0.077)	Loss 1.021 (0.900)
Epoch: [15][140/200]	Time 0.269 (0.349)	Data 0.000 (0.076)	Loss 1.298 (0.926)
Epoch: [15][160/200]	Time 0.269 (0.348)	Data 0.000 (0.075)	Loss 0.977 (0.939)
Epoch: [15][180/200]	Time 0.264 (0.348)	Data 0.000 (0.075)	Loss 1.329 (0.953)
Epoch: [15][200/200]	Time 0.267 (0.355)	Data 0.001 (0.082)	Loss 1.005 (0.962)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.138)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.788983821868896
==> Statistics for epoch 16: 593 clusters
Epoch: [16][20/200]	Time 0.268 (0.387)	Data 0.001 (0.110)	Loss 0.884 (0.225)
Epoch: [16][40/200]	Time 0.265 (0.364)	Data 0.001 (0.092)	Loss 0.880 (0.594)
Epoch: [16][60/200]	Time 0.269 (0.359)	Data 0.001 (0.086)	Loss 1.103 (0.746)
Epoch: [16][80/200]	Time 0.266 (0.355)	Data 0.001 (0.082)	Loss 0.892 (0.813)
Epoch: [16][100/200]	Time 0.268 (0.353)	Data 0.001 (0.080)	Loss 1.078 (0.843)
Epoch: [16][120/200]	Time 0.265 (0.352)	Data 0.000 (0.079)	Loss 1.197 (0.875)
Epoch: [16][140/200]	Time 0.268 (0.351)	Data 0.000 (0.077)	Loss 1.117 (0.898)
Epoch: [16][160/200]	Time 0.268 (0.351)	Data 0.000 (0.077)	Loss 1.179 (0.909)
Epoch: [16][180/200]	Time 0.269 (0.350)	Data 0.000 (0.076)	Loss 1.071 (0.917)
Epoch: [16][200/200]	Time 0.269 (0.357)	Data 0.001 (0.083)	Loss 0.658 (0.925)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.103 (0.139)	Data 0.006 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.196364164352417
==> Statistics for epoch 17: 593 clusters
Epoch: [17][20/200]	Time 0.272 (0.379)	Data 0.001 (0.109)	Loss 1.056 (0.224)
Epoch: [17][40/200]	Time 0.267 (0.365)	Data 0.001 (0.092)	Loss 0.730 (0.561)
Epoch: [17][60/200]	Time 0.266 (0.358)	Data 0.001 (0.084)	Loss 1.057 (0.706)
Epoch: [17][80/200]	Time 0.264 (0.353)	Data 0.000 (0.080)	Loss 1.485 (0.783)
Epoch: [17][100/200]	Time 0.266 (0.350)	Data 0.000 (0.077)	Loss 1.381 (0.824)
Epoch: [17][120/200]	Time 0.267 (0.350)	Data 0.001 (0.078)	Loss 0.878 (0.848)
Epoch: [17][140/200]	Time 0.267 (0.350)	Data 0.000 (0.077)	Loss 1.071 (0.876)
Epoch: [17][160/200]	Time 0.269 (0.350)	Data 0.000 (0.077)	Loss 0.714 (0.894)
Epoch: [17][180/200]	Time 0.268 (0.349)	Data 0.000 (0.077)	Loss 0.927 (0.913)
Epoch: [17][200/200]	Time 0.274 (0.356)	Data 0.000 (0.083)	Loss 0.895 (0.924)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.098 (0.137)	Data 0.000 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.541034936904907
==> Statistics for epoch 18: 594 clusters
Epoch: [18][20/200]	Time 0.263 (0.379)	Data 0.001 (0.105)	Loss 0.914 (0.216)
Epoch: [18][40/200]	Time 0.268 (0.359)	Data 0.001 (0.088)	Loss 0.976 (0.591)
Epoch: [18][60/200]	Time 0.267 (0.353)	Data 0.001 (0.083)	Loss 0.964 (0.728)
Epoch: [18][80/200]	Time 0.269 (0.351)	Data 0.001 (0.080)	Loss 0.827 (0.772)
Epoch: [18][100/200]	Time 0.265 (0.351)	Data 0.001 (0.079)	Loss 1.272 (0.822)
Epoch: [18][120/200]	Time 0.267 (0.351)	Data 0.000 (0.079)	Loss 1.099 (0.842)
Epoch: [18][140/200]	Time 0.267 (0.349)	Data 0.000 (0.078)	Loss 0.648 (0.855)
Epoch: [18][160/200]	Time 0.265 (0.349)	Data 0.000 (0.077)	Loss 1.002 (0.874)
Epoch: [18][180/200]	Time 0.267 (0.349)	Data 0.000 (0.077)	Loss 1.043 (0.881)
Epoch: [18][200/200]	Time 0.270 (0.356)	Data 0.001 (0.085)	Loss 0.899 (0.888)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.131)	Data 0.000 (0.030)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.745557069778442
==> Statistics for epoch 19: 593 clusters
Epoch: [19][20/200]	Time 0.275 (0.382)	Data 0.001 (0.105)	Loss 0.846 (0.214)
Epoch: [19][40/200]	Time 0.382 (0.363)	Data 0.001 (0.087)	Loss 0.884 (0.551)
Epoch: [19][60/200]	Time 0.264 (0.354)	Data 0.001 (0.081)	Loss 1.172 (0.664)
Epoch: [19][80/200]	Time 0.268 (0.348)	Data 0.001 (0.076)	Loss 0.958 (0.725)
Epoch: [19][100/200]	Time 0.267 (0.346)	Data 0.001 (0.075)	Loss 0.853 (0.769)
Epoch: [19][120/200]	Time 0.267 (0.345)	Data 0.000 (0.073)	Loss 0.545 (0.787)
Epoch: [19][140/200]	Time 0.351 (0.344)	Data 0.000 (0.072)	Loss 0.878 (0.807)
Epoch: [19][160/200]	Time 0.264 (0.343)	Data 0.000 (0.071)	Loss 1.033 (0.828)
Epoch: [19][180/200]	Time 0.264 (0.343)	Data 0.000 (0.071)	Loss 0.755 (0.840)
Epoch: [19][200/200]	Time 0.267 (0.349)	Data 0.001 (0.077)	Loss 0.790 (0.847)
Extract Features: [50/76]	Time 0.195 (0.138)	Data 0.097 (0.037)	
Mean AP: 90.2%

 * Finished epoch  19  model mAP: 90.2%  best: 90.2% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.140)	Data 0.000 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.812493562698364
==> Statistics for epoch 20: 593 clusters
Epoch: [20][20/200]	Time 0.270 (0.381)	Data 0.001 (0.106)	Loss 0.630 (0.183)
Epoch: [20][40/200]	Time 0.269 (0.359)	Data 0.001 (0.085)	Loss 0.946 (0.547)
Epoch: [20][60/200]	Time 0.265 (0.351)	Data 0.001 (0.079)	Loss 1.384 (0.666)
Epoch: [20][80/200]	Time 0.266 (0.348)	Data 0.001 (0.075)	Loss 0.636 (0.711)
Epoch: [20][100/200]	Time 0.267 (0.347)	Data 0.001 (0.074)	Loss 1.034 (0.746)
Epoch: [20][120/200]	Time 0.266 (0.346)	Data 0.000 (0.073)	Loss 1.249 (0.765)
Epoch: [20][140/200]	Time 0.263 (0.344)	Data 0.000 (0.072)	Loss 0.890 (0.778)
Epoch: [20][160/200]	Time 0.266 (0.345)	Data 0.000 (0.072)	Loss 1.175 (0.796)
Epoch: [20][180/200]	Time 0.267 (0.344)	Data 0.000 (0.071)	Loss 0.948 (0.800)
Epoch: [20][200/200]	Time 0.264 (0.350)	Data 0.001 (0.077)	Loss 0.907 (0.810)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.098 (0.140)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.39851450920105
==> Statistics for epoch 21: 594 clusters
Epoch: [21][20/200]	Time 0.275 (0.379)	Data 0.001 (0.104)	Loss 0.570 (0.192)
Epoch: [21][40/200]	Time 0.269 (0.358)	Data 0.001 (0.087)	Loss 0.946 (0.505)
Epoch: [21][60/200]	Time 0.271 (0.350)	Data 0.001 (0.078)	Loss 0.900 (0.631)
Epoch: [21][80/200]	Time 0.267 (0.348)	Data 0.001 (0.074)	Loss 0.540 (0.679)
Epoch: [21][100/200]	Time 0.265 (0.347)	Data 0.001 (0.073)	Loss 1.048 (0.712)
Epoch: [21][120/200]	Time 0.264 (0.345)	Data 0.000 (0.072)	Loss 0.627 (0.731)
Epoch: [21][140/200]	Time 0.266 (0.346)	Data 0.000 (0.073)	Loss 1.191 (0.753)
Epoch: [21][160/200]	Time 0.395 (0.346)	Data 0.000 (0.072)	Loss 0.733 (0.762)
Epoch: [21][180/200]	Time 0.265 (0.345)	Data 0.000 (0.072)	Loss 1.027 (0.780)
Epoch: [21][200/200]	Time 0.270 (0.351)	Data 0.001 (0.078)	Loss 0.608 (0.791)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.135)	Data 0.000 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.488036632537842
==> Statistics for epoch 22: 594 clusters
Epoch: [22][20/200]	Time 0.269 (0.381)	Data 0.000 (0.103)	Loss 0.855 (0.192)
Epoch: [22][40/200]	Time 0.270 (0.359)	Data 0.000 (0.083)	Loss 0.848 (0.513)
Epoch: [22][60/200]	Time 0.264 (0.352)	Data 0.000 (0.077)	Loss 0.704 (0.615)
Epoch: [22][80/200]	Time 0.263 (0.349)	Data 0.000 (0.076)	Loss 0.682 (0.664)
Epoch: [22][100/200]	Time 0.268 (0.348)	Data 0.001 (0.074)	Loss 0.937 (0.716)
Epoch: [22][120/200]	Time 0.266 (0.347)	Data 0.000 (0.073)	Loss 0.822 (0.732)
Epoch: [22][140/200]	Time 0.264 (0.345)	Data 0.000 (0.072)	Loss 1.088 (0.750)
Epoch: [22][160/200]	Time 0.266 (0.344)	Data 0.000 (0.071)	Loss 0.924 (0.760)
Epoch: [22][180/200]	Time 0.269 (0.343)	Data 0.000 (0.070)	Loss 1.281 (0.775)
Epoch: [22][200/200]	Time 0.271 (0.349)	Data 0.000 (0.076)	Loss 0.798 (0.787)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.131)	Data 0.000 (0.030)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.00823450088501
==> Statistics for epoch 23: 593 clusters
Epoch: [23][20/200]	Time 0.270 (0.378)	Data 0.001 (0.104)	Loss 0.751 (0.200)
Epoch: [23][40/200]	Time 0.270 (0.362)	Data 0.001 (0.087)	Loss 0.950 (0.526)
Epoch: [23][60/200]	Time 0.266 (0.353)	Data 0.001 (0.079)	Loss 0.885 (0.628)
Epoch: [23][80/200]	Time 0.266 (0.353)	Data 0.000 (0.079)	Loss 0.842 (0.690)
Epoch: [23][100/200]	Time 0.263 (0.353)	Data 0.001 (0.079)	Loss 0.748 (0.719)
Epoch: [23][120/200]	Time 0.267 (0.351)	Data 0.000 (0.078)	Loss 0.610 (0.741)
Epoch: [23][140/200]	Time 0.269 (0.350)	Data 0.000 (0.077)	Loss 1.098 (0.763)
Epoch: [23][160/200]	Time 0.265 (0.349)	Data 0.000 (0.075)	Loss 0.900 (0.768)
Epoch: [23][180/200]	Time 0.267 (0.348)	Data 0.000 (0.074)	Loss 1.024 (0.776)
Epoch: [23][200/200]	Time 0.269 (0.354)	Data 0.001 (0.080)	Loss 0.825 (0.783)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.137)	Data 0.000 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.858914613723755
==> Statistics for epoch 24: 592 clusters
Epoch: [24][20/200]	Time 0.281 (0.391)	Data 0.001 (0.115)	Loss 0.692 (0.191)
Epoch: [24][40/200]	Time 0.268 (0.371)	Data 0.001 (0.097)	Loss 0.891 (0.522)
Epoch: [24][60/200]	Time 0.267 (0.359)	Data 0.001 (0.087)	Loss 0.924 (0.649)
Epoch: [24][80/200]	Time 0.271 (0.355)	Data 0.001 (0.083)	Loss 0.729 (0.705)
Epoch: [24][100/200]	Time 0.265 (0.351)	Data 0.001 (0.079)	Loss 0.648 (0.738)
Epoch: [24][120/200]	Time 0.267 (0.350)	Data 0.000 (0.077)	Loss 0.909 (0.757)
Epoch: [24][140/200]	Time 0.263 (0.350)	Data 0.000 (0.077)	Loss 0.584 (0.769)
Epoch: [24][160/200]	Time 0.262 (0.349)	Data 0.000 (0.076)	Loss 0.753 (0.773)
Epoch: [24][180/200]	Time 0.265 (0.348)	Data 0.000 (0.075)	Loss 1.093 (0.780)
Epoch: [24][200/200]	Time 0.266 (0.354)	Data 0.001 (0.081)	Loss 0.728 (0.786)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.132)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.019725561141968
==> Statistics for epoch 25: 592 clusters
Epoch: [25][20/200]	Time 0.268 (0.372)	Data 0.001 (0.101)	Loss 0.806 (0.165)
Epoch: [25][40/200]	Time 0.270 (0.357)	Data 0.001 (0.085)	Loss 0.900 (0.461)
Epoch: [25][60/200]	Time 0.279 (0.353)	Data 0.002 (0.080)	Loss 1.049 (0.559)
Epoch: [25][80/200]	Time 0.265 (0.351)	Data 0.000 (0.077)	Loss 0.865 (0.624)
Epoch: [25][100/200]	Time 0.268 (0.348)	Data 0.001 (0.075)	Loss 0.882 (0.650)
Epoch: [25][120/200]	Time 0.266 (0.348)	Data 0.000 (0.074)	Loss 1.108 (0.678)
Epoch: [25][140/200]	Time 0.263 (0.346)	Data 0.000 (0.072)	Loss 0.959 (0.704)
Epoch: [25][160/200]	Time 0.269 (0.345)	Data 0.000 (0.071)	Loss 1.387 (0.723)
Epoch: [25][180/200]	Time 0.266 (0.344)	Data 0.000 (0.070)	Loss 0.903 (0.737)
Epoch: [25][200/200]	Time 0.273 (0.351)	Data 0.001 (0.077)	Loss 0.859 (0.750)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.098 (0.133)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.887116193771362
==> Statistics for epoch 26: 595 clusters
Epoch: [26][20/200]	Time 0.270 (0.372)	Data 0.001 (0.097)	Loss 0.748 (0.175)
Epoch: [26][40/200]	Time 0.267 (0.360)	Data 0.001 (0.086)	Loss 1.001 (0.489)
Epoch: [26][60/200]	Time 0.275 (0.357)	Data 0.001 (0.083)	Loss 0.745 (0.598)
Epoch: [26][80/200]	Time 0.338 (0.354)	Data 0.001 (0.081)	Loss 0.952 (0.659)
Epoch: [26][100/200]	Time 0.266 (0.352)	Data 0.001 (0.080)	Loss 0.716 (0.696)
Epoch: [26][120/200]	Time 0.265 (0.351)	Data 0.000 (0.078)	Loss 0.898 (0.721)
Epoch: [26][140/200]	Time 0.268 (0.350)	Data 0.000 (0.078)	Loss 0.865 (0.743)
Epoch: [26][160/200]	Time 0.268 (0.350)	Data 0.000 (0.078)	Loss 0.796 (0.751)
Epoch: [26][180/200]	Time 0.267 (0.350)	Data 0.000 (0.078)	Loss 0.592 (0.770)
Epoch: [26][200/200]	Time 0.269 (0.357)	Data 0.001 (0.084)	Loss 0.940 (0.776)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.139)	Data 0.000 (0.040)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.339917182922363
==> Statistics for epoch 27: 592 clusters
Epoch: [27][20/200]	Time 0.273 (0.383)	Data 0.001 (0.113)	Loss 0.913 (0.180)
Epoch: [27][40/200]	Time 0.269 (0.364)	Data 0.001 (0.091)	Loss 1.143 (0.487)
Epoch: [27][60/200]	Time 0.269 (0.357)	Data 0.001 (0.083)	Loss 1.142 (0.582)
Epoch: [27][80/200]	Time 0.386 (0.354)	Data 0.001 (0.080)	Loss 0.780 (0.645)
Epoch: [27][100/200]	Time 0.267 (0.351)	Data 0.001 (0.078)	Loss 0.953 (0.678)
Epoch: [27][120/200]	Time 0.268 (0.350)	Data 0.000 (0.077)	Loss 0.638 (0.704)
Epoch: [27][140/200]	Time 0.266 (0.349)	Data 0.000 (0.076)	Loss 0.730 (0.726)
Epoch: [27][160/200]	Time 0.273 (0.349)	Data 0.000 (0.076)	Loss 0.755 (0.740)
Epoch: [27][180/200]	Time 0.267 (0.349)	Data 0.000 (0.075)	Loss 0.912 (0.746)
Epoch: [27][200/200]	Time 0.273 (0.356)	Data 0.001 (0.083)	Loss 0.865 (0.753)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.143)	Data 0.000 (0.043)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.51943612098694
==> Statistics for epoch 28: 592 clusters
Epoch: [28][20/200]	Time 0.269 (0.379)	Data 0.001 (0.108)	Loss 0.844 (0.168)
Epoch: [28][40/200]	Time 0.268 (0.364)	Data 0.001 (0.092)	Loss 1.116 (0.471)
Epoch: [28][60/200]	Time 0.266 (0.358)	Data 0.001 (0.085)	Loss 0.849 (0.589)
Epoch: [28][80/200]	Time 0.267 (0.354)	Data 0.001 (0.082)	Loss 0.708 (0.649)
Epoch: [28][100/200]	Time 0.266 (0.352)	Data 0.000 (0.080)	Loss 0.909 (0.691)
Epoch: [28][120/200]	Time 0.267 (0.351)	Data 0.000 (0.079)	Loss 0.678 (0.710)
Epoch: [28][140/200]	Time 0.268 (0.350)	Data 0.000 (0.077)	Loss 0.716 (0.723)
Epoch: [28][160/200]	Time 0.266 (0.350)	Data 0.000 (0.077)	Loss 0.949 (0.739)
Epoch: [28][180/200]	Time 0.268 (0.350)	Data 0.000 (0.077)	Loss 0.888 (0.746)
Epoch: [28][200/200]	Time 0.267 (0.358)	Data 0.001 (0.084)	Loss 0.834 (0.756)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.143)	Data 0.000 (0.041)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.85444951057434
==> Statistics for epoch 29: 592 clusters
Epoch: [29][20/200]	Time 0.272 (0.374)	Data 0.001 (0.097)	Loss 0.778 (0.177)
Epoch: [29][40/200]	Time 0.265 (0.357)	Data 0.001 (0.082)	Loss 0.629 (0.483)
Epoch: [29][60/200]	Time 0.270 (0.349)	Data 0.001 (0.077)	Loss 0.655 (0.595)
Epoch: [29][80/200]	Time 0.265 (0.346)	Data 0.001 (0.074)	Loss 1.125 (0.663)
Epoch: [29][100/200]	Time 0.269 (0.345)	Data 0.001 (0.073)	Loss 0.735 (0.689)
Epoch: [29][120/200]	Time 0.267 (0.344)	Data 0.000 (0.072)	Loss 0.762 (0.707)
Epoch: [29][140/200]	Time 0.266 (0.344)	Data 0.000 (0.072)	Loss 0.663 (0.731)
Epoch: [29][160/200]	Time 0.267 (0.343)	Data 0.000 (0.072)	Loss 0.971 (0.746)
Epoch: [29][180/200]	Time 0.265 (0.342)	Data 0.000 (0.070)	Loss 0.716 (0.758)
Epoch: [29][200/200]	Time 0.268 (0.349)	Data 0.000 (0.077)	Loss 0.682 (0.768)
Extract Features: [50/76]	Time 0.178 (0.142)	Data 0.002 (0.041)	
Mean AP: 90.3%

 * Finished epoch  29  model mAP: 90.3%  best: 90.3% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.141)	Data 0.000 (0.041)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.515581607818604
==> Statistics for epoch 30: 592 clusters
Epoch: [30][20/200]	Time 0.267 (0.366)	Data 0.001 (0.096)	Loss 0.960 (0.184)
Epoch: [30][40/200]	Time 0.265 (0.356)	Data 0.001 (0.083)	Loss 0.718 (0.452)
Epoch: [30][60/200]	Time 0.267 (0.350)	Data 0.001 (0.077)	Loss 0.966 (0.579)
Epoch: [30][80/200]	Time 0.267 (0.346)	Data 0.001 (0.074)	Loss 0.907 (0.637)
Epoch: [30][100/200]	Time 0.264 (0.345)	Data 0.001 (0.072)	Loss 0.746 (0.685)
Epoch: [30][120/200]	Time 0.268 (0.344)	Data 0.000 (0.070)	Loss 1.320 (0.710)
Epoch: [30][140/200]	Time 0.264 (0.345)	Data 0.000 (0.071)	Loss 0.844 (0.724)
Epoch: [30][160/200]	Time 0.268 (0.343)	Data 0.000 (0.070)	Loss 0.642 (0.731)
Epoch: [30][180/200]	Time 0.265 (0.343)	Data 0.000 (0.070)	Loss 1.033 (0.743)
Epoch: [30][200/200]	Time 0.267 (0.349)	Data 0.001 (0.076)	Loss 0.852 (0.752)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.134)	Data 0.000 (0.033)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.244102478027344
==> Statistics for epoch 31: 595 clusters
Epoch: [31][20/200]	Time 0.277 (0.378)	Data 0.001 (0.101)	Loss 0.688 (0.165)
Epoch: [31][40/200]	Time 0.364 (0.356)	Data 0.001 (0.082)	Loss 0.832 (0.486)
Epoch: [31][60/200]	Time 0.267 (0.348)	Data 0.001 (0.076)	Loss 0.713 (0.586)
Epoch: [31][80/200]	Time 0.266 (0.345)	Data 0.001 (0.073)	Loss 0.708 (0.648)
Epoch: [31][100/200]	Time 0.265 (0.343)	Data 0.001 (0.070)	Loss 0.859 (0.684)
Epoch: [31][120/200]	Time 0.268 (0.342)	Data 0.000 (0.070)	Loss 0.997 (0.714)
Epoch: [31][140/200]	Time 0.265 (0.341)	Data 0.000 (0.069)	Loss 0.905 (0.725)
Epoch: [31][160/200]	Time 0.265 (0.341)	Data 0.000 (0.069)	Loss 1.134 (0.732)
Epoch: [31][180/200]	Time 0.266 (0.340)	Data 0.000 (0.069)	Loss 0.969 (0.741)
Epoch: [31][200/200]	Time 0.278 (0.348)	Data 0.001 (0.075)	Loss 0.932 (0.753)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.133)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.202332258224487
==> Statistics for epoch 32: 594 clusters
Epoch: [32][20/200]	Time 0.269 (0.376)	Data 0.001 (0.106)	Loss 0.696 (0.163)
Epoch: [32][40/200]	Time 0.271 (0.358)	Data 0.001 (0.087)	Loss 1.052 (0.490)
Epoch: [32][60/200]	Time 0.267 (0.355)	Data 0.001 (0.082)	Loss 0.737 (0.610)
Epoch: [32][80/200]	Time 0.266 (0.349)	Data 0.001 (0.078)	Loss 0.482 (0.654)
Epoch: [32][100/200]	Time 0.268 (0.348)	Data 0.001 (0.076)	Loss 0.643 (0.690)
Epoch: [32][120/200]	Time 0.263 (0.346)	Data 0.000 (0.074)	Loss 0.938 (0.711)
Epoch: [32][140/200]	Time 0.261 (0.344)	Data 0.000 (0.072)	Loss 0.762 (0.721)
Epoch: [32][160/200]	Time 0.264 (0.344)	Data 0.000 (0.072)	Loss 1.339 (0.741)
Epoch: [32][180/200]	Time 0.262 (0.343)	Data 0.000 (0.071)	Loss 0.775 (0.755)
Epoch: [32][200/200]	Time 0.280 (0.349)	Data 0.001 (0.078)	Loss 0.625 (0.763)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.171 (0.136)	Data 0.075 (0.034)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.516452074050903
==> Statistics for epoch 33: 596 clusters
Epoch: [33][20/200]	Time 0.281 (0.380)	Data 0.001 (0.103)	Loss 0.859 (0.179)
Epoch: [33][40/200]	Time 0.265 (0.360)	Data 0.001 (0.086)	Loss 0.624 (0.470)
Epoch: [33][60/200]	Time 0.269 (0.354)	Data 0.001 (0.080)	Loss 0.733 (0.570)
Epoch: [33][80/200]	Time 0.269 (0.348)	Data 0.001 (0.076)	Loss 0.549 (0.630)
Epoch: [33][100/200]	Time 0.271 (0.347)	Data 0.001 (0.074)	Loss 0.678 (0.668)
Epoch: [33][120/200]	Time 0.269 (0.347)	Data 0.000 (0.074)	Loss 1.129 (0.694)
Epoch: [33][140/200]	Time 0.268 (0.346)	Data 0.000 (0.072)	Loss 0.816 (0.714)
Epoch: [33][160/200]	Time 0.266 (0.345)	Data 0.000 (0.072)	Loss 0.587 (0.726)
Epoch: [33][180/200]	Time 0.265 (0.345)	Data 0.000 (0.072)	Loss 0.605 (0.733)
Epoch: [33][200/200]	Time 0.268 (0.353)	Data 0.001 (0.079)	Loss 0.521 (0.749)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.138)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.140029191970825
==> Statistics for epoch 34: 594 clusters
Epoch: [34][20/200]	Time 0.266 (0.377)	Data 0.001 (0.098)	Loss 0.670 (0.188)
Epoch: [34][40/200]	Time 0.270 (0.357)	Data 0.001 (0.080)	Loss 1.107 (0.507)
Epoch: [34][60/200]	Time 0.269 (0.353)	Data 0.001 (0.077)	Loss 0.774 (0.628)
Epoch: [34][80/200]	Time 0.266 (0.348)	Data 0.001 (0.073)	Loss 0.840 (0.686)
Epoch: [34][100/200]	Time 0.266 (0.349)	Data 0.001 (0.075)	Loss 0.811 (0.723)
Epoch: [34][120/200]	Time 0.273 (0.348)	Data 0.000 (0.074)	Loss 0.894 (0.738)
Epoch: [34][140/200]	Time 0.265 (0.348)	Data 0.000 (0.074)	Loss 0.657 (0.756)
Epoch: [34][160/200]	Time 0.263 (0.346)	Data 0.000 (0.073)	Loss 0.828 (0.767)
Epoch: [34][180/200]	Time 0.266 (0.345)	Data 0.000 (0.072)	Loss 1.067 (0.767)
Epoch: [34][200/200]	Time 0.269 (0.351)	Data 0.001 (0.078)	Loss 0.986 (0.773)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.136)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.317524194717407
==> Statistics for epoch 35: 595 clusters
Epoch: [35][20/200]	Time 0.267 (0.378)	Data 0.001 (0.101)	Loss 0.991 (0.186)
Epoch: [35][40/200]	Time 0.274 (0.361)	Data 0.001 (0.089)	Loss 0.984 (0.463)
Epoch: [35][60/200]	Time 0.267 (0.354)	Data 0.001 (0.081)	Loss 0.946 (0.600)
Epoch: [35][80/200]	Time 0.265 (0.349)	Data 0.001 (0.078)	Loss 0.810 (0.643)
Epoch: [35][100/200]	Time 0.266 (0.348)	Data 0.001 (0.076)	Loss 0.985 (0.673)
Epoch: [35][120/200]	Time 0.265 (0.347)	Data 0.000 (0.075)	Loss 0.767 (0.691)
Epoch: [35][140/200]	Time 0.264 (0.346)	Data 0.000 (0.074)	Loss 0.729 (0.705)
Epoch: [35][160/200]	Time 0.268 (0.344)	Data 0.000 (0.073)	Loss 0.479 (0.713)
Epoch: [35][180/200]	Time 0.264 (0.344)	Data 0.000 (0.073)	Loss 0.862 (0.721)
Epoch: [35][200/200]	Time 0.270 (0.352)	Data 0.001 (0.079)	Loss 0.836 (0.734)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.142)	Data 0.000 (0.041)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.919614791870117
==> Statistics for epoch 36: 593 clusters
Epoch: [36][20/200]	Time 0.264 (0.377)	Data 0.001 (0.107)	Loss 0.906 (0.197)
Epoch: [36][40/200]	Time 0.270 (0.364)	Data 0.001 (0.092)	Loss 0.964 (0.499)
Epoch: [36][60/200]	Time 0.265 (0.356)	Data 0.001 (0.083)	Loss 0.711 (0.594)
Epoch: [36][80/200]	Time 0.265 (0.352)	Data 0.001 (0.080)	Loss 0.770 (0.651)
Epoch: [36][100/200]	Time 0.268 (0.350)	Data 0.001 (0.078)	Loss 0.824 (0.676)
Epoch: [36][120/200]	Time 0.268 (0.348)	Data 0.000 (0.076)	Loss 0.983 (0.693)
Epoch: [36][140/200]	Time 0.269 (0.347)	Data 0.000 (0.076)	Loss 0.859 (0.705)
Epoch: [36][160/200]	Time 0.268 (0.348)	Data 0.000 (0.076)	Loss 0.774 (0.712)
Epoch: [36][180/200]	Time 0.263 (0.347)	Data 0.000 (0.075)	Loss 0.630 (0.725)
Epoch: [36][200/200]	Time 0.279 (0.354)	Data 0.001 (0.082)	Loss 0.595 (0.733)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.139)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.904911041259766
==> Statistics for epoch 37: 594 clusters
Epoch: [37][20/200]	Time 0.265 (0.393)	Data 0.001 (0.117)	Loss 0.842 (0.164)
Epoch: [37][40/200]	Time 0.270 (0.369)	Data 0.001 (0.097)	Loss 0.855 (0.464)
Epoch: [37][60/200]	Time 0.267 (0.362)	Data 0.001 (0.089)	Loss 0.750 (0.573)
Epoch: [37][80/200]	Time 0.269 (0.359)	Data 0.001 (0.086)	Loss 0.679 (0.628)
Epoch: [37][100/200]	Time 0.263 (0.357)	Data 0.001 (0.084)	Loss 0.824 (0.660)
Epoch: [37][120/200]	Time 0.270 (0.354)	Data 0.000 (0.082)	Loss 0.921 (0.685)
Epoch: [37][140/200]	Time 0.267 (0.353)	Data 0.000 (0.081)	Loss 0.805 (0.701)
Epoch: [37][160/200]	Time 0.269 (0.352)	Data 0.000 (0.080)	Loss 0.857 (0.705)
Epoch: [37][180/200]	Time 0.273 (0.352)	Data 0.000 (0.080)	Loss 1.322 (0.718)
Epoch: [37][200/200]	Time 0.272 (0.358)	Data 0.001 (0.086)	Loss 0.674 (0.726)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.141)	Data 0.000 (0.036)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.46994113922119
==> Statistics for epoch 38: 594 clusters
Epoch: [38][20/200]	Time 0.269 (0.378)	Data 0.001 (0.100)	Loss 0.831 (0.175)
Epoch: [38][40/200]	Time 0.270 (0.362)	Data 0.001 (0.086)	Loss 0.531 (0.470)
Epoch: [38][60/200]	Time 0.291 (0.356)	Data 0.001 (0.083)	Loss 0.905 (0.604)
Epoch: [38][80/200]	Time 0.264 (0.353)	Data 0.001 (0.081)	Loss 0.845 (0.645)
Epoch: [38][100/200]	Time 0.270 (0.352)	Data 0.001 (0.079)	Loss 0.834 (0.683)
Epoch: [38][120/200]	Time 0.267 (0.350)	Data 0.000 (0.078)	Loss 1.006 (0.705)
Epoch: [38][140/200]	Time 0.269 (0.350)	Data 0.000 (0.078)	Loss 0.925 (0.722)
Epoch: [38][160/200]	Time 0.267 (0.349)	Data 0.000 (0.077)	Loss 1.007 (0.732)
Epoch: [38][180/200]	Time 0.268 (0.349)	Data 0.000 (0.077)	Loss 0.485 (0.739)
Epoch: [38][200/200]	Time 0.274 (0.356)	Data 0.001 (0.084)	Loss 0.703 (0.744)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.137)	Data 0.000 (0.036)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.23146414756775
==> Statistics for epoch 39: 595 clusters
Epoch: [39][20/200]	Time 0.269 (0.378)	Data 0.001 (0.107)	Loss 0.587 (0.158)
Epoch: [39][40/200]	Time 0.274 (0.362)	Data 0.001 (0.091)	Loss 0.751 (0.452)
Epoch: [39][60/200]	Time 0.266 (0.358)	Data 0.001 (0.086)	Loss 0.735 (0.582)
Epoch: [39][80/200]	Time 0.268 (0.356)	Data 0.001 (0.084)	Loss 0.718 (0.624)
Epoch: [39][100/200]	Time 0.267 (0.354)	Data 0.001 (0.082)	Loss 1.029 (0.662)
Epoch: [39][120/200]	Time 0.269 (0.352)	Data 0.000 (0.080)	Loss 0.659 (0.680)
Epoch: [39][140/200]	Time 0.269 (0.351)	Data 0.000 (0.078)	Loss 0.893 (0.694)
Epoch: [39][160/200]	Time 0.269 (0.351)	Data 0.000 (0.078)	Loss 1.027 (0.708)
Epoch: [39][180/200]	Time 0.264 (0.350)	Data 0.000 (0.077)	Loss 0.598 (0.723)
Epoch: [39][200/200]	Time 0.276 (0.357)	Data 0.001 (0.084)	Loss 0.984 (0.730)
Extract Features: [50/76]	Time 0.139 (0.138)	Data 0.042 (0.036)	
Mean AP: 90.3%

 * Finished epoch  39  model mAP: 90.3%  best: 90.3%

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.140)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.44965887069702
==> Statistics for epoch 40: 595 clusters
Epoch: [40][20/200]	Time 0.270 (0.376)	Data 0.001 (0.101)	Loss 0.899 (0.171)
Epoch: [40][40/200]	Time 0.267 (0.363)	Data 0.001 (0.089)	Loss 0.604 (0.469)
Epoch: [40][60/200]	Time 0.268 (0.356)	Data 0.001 (0.083)	Loss 0.678 (0.569)
Epoch: [40][80/200]	Time 0.266 (0.354)	Data 0.001 (0.082)	Loss 0.737 (0.624)
Epoch: [40][100/200]	Time 0.267 (0.351)	Data 0.001 (0.079)	Loss 1.001 (0.654)
Epoch: [40][120/200]	Time 0.265 (0.350)	Data 0.000 (0.077)	Loss 0.849 (0.673)
Epoch: [40][140/200]	Time 0.268 (0.349)	Data 0.000 (0.077)	Loss 0.820 (0.689)
Epoch: [40][160/200]	Time 0.267 (0.349)	Data 0.000 (0.077)	Loss 1.108 (0.709)
Epoch: [40][180/200]	Time 0.267 (0.349)	Data 0.000 (0.076)	Loss 0.659 (0.716)
Epoch: [40][200/200]	Time 0.268 (0.356)	Data 0.001 (0.083)	Loss 0.825 (0.722)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.145)	Data 0.000 (0.044)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.555591344833374
==> Statistics for epoch 41: 593 clusters
Epoch: [41][20/200]	Time 0.269 (0.385)	Data 0.001 (0.108)	Loss 0.684 (0.159)
Epoch: [41][40/200]	Time 0.270 (0.366)	Data 0.001 (0.089)	Loss 0.761 (0.442)
Epoch: [41][60/200]	Time 0.270 (0.360)	Data 0.001 (0.084)	Loss 0.767 (0.551)
Epoch: [41][80/200]	Time 0.265 (0.356)	Data 0.000 (0.082)	Loss 0.801 (0.615)
Epoch: [41][100/200]	Time 0.269 (0.354)	Data 0.001 (0.080)	Loss 0.710 (0.646)
Epoch: [41][120/200]	Time 0.267 (0.352)	Data 0.000 (0.078)	Loss 0.541 (0.673)
Epoch: [41][140/200]	Time 0.264 (0.351)	Data 0.000 (0.077)	Loss 0.763 (0.685)
Epoch: [41][160/200]	Time 0.269 (0.351)	Data 0.000 (0.077)	Loss 0.734 (0.702)
Epoch: [41][180/200]	Time 0.267 (0.350)	Data 0.000 (0.076)	Loss 1.277 (0.725)
Epoch: [41][200/200]	Time 0.272 (0.356)	Data 0.000 (0.082)	Loss 0.864 (0.734)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.175 (0.136)	Data 0.000 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.078317403793335
==> Statistics for epoch 42: 593 clusters
Epoch: [42][20/200]	Time 0.275 (0.385)	Data 0.001 (0.109)	Loss 0.662 (0.157)
Epoch: [42][40/200]	Time 0.276 (0.361)	Data 0.001 (0.089)	Loss 1.006 (0.467)
Epoch: [42][60/200]	Time 0.269 (0.356)	Data 0.001 (0.083)	Loss 1.087 (0.577)
Epoch: [42][80/200]	Time 0.269 (0.354)	Data 0.001 (0.080)	Loss 0.664 (0.629)
Epoch: [42][100/200]	Time 0.270 (0.351)	Data 0.001 (0.078)	Loss 0.660 (0.660)
Epoch: [42][120/200]	Time 0.268 (0.350)	Data 0.000 (0.077)	Loss 0.803 (0.690)
Epoch: [42][140/200]	Time 0.269 (0.350)	Data 0.000 (0.077)	Loss 0.678 (0.698)
Epoch: [42][160/200]	Time 0.270 (0.350)	Data 0.000 (0.076)	Loss 0.618 (0.705)
Epoch: [42][180/200]	Time 0.269 (0.348)	Data 0.000 (0.076)	Loss 0.598 (0.713)
Epoch: [42][200/200]	Time 0.273 (0.356)	Data 0.001 (0.083)	Loss 0.681 (0.719)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.139)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.48276710510254
==> Statistics for epoch 43: 593 clusters
Epoch: [43][20/200]	Time 0.268 (0.384)	Data 0.001 (0.115)	Loss 0.752 (0.165)
Epoch: [43][40/200]	Time 0.266 (0.367)	Data 0.001 (0.095)	Loss 0.574 (0.467)
Epoch: [43][60/200]	Time 0.268 (0.361)	Data 0.001 (0.088)	Loss 0.607 (0.581)
Epoch: [43][80/200]	Time 0.269 (0.355)	Data 0.001 (0.083)	Loss 0.750 (0.641)
Epoch: [43][100/200]	Time 0.268 (0.353)	Data 0.001 (0.081)	Loss 0.646 (0.662)
Epoch: [43][120/200]	Time 0.267 (0.352)	Data 0.000 (0.079)	Loss 0.619 (0.686)
Epoch: [43][140/200]	Time 0.268 (0.352)	Data 0.000 (0.079)	Loss 0.663 (0.708)
Epoch: [43][160/200]	Time 0.270 (0.350)	Data 0.000 (0.078)	Loss 0.564 (0.717)
Epoch: [43][180/200]	Time 0.266 (0.350)	Data 0.000 (0.077)	Loss 0.710 (0.725)
Epoch: [43][200/200]	Time 0.270 (0.357)	Data 0.001 (0.084)	Loss 0.606 (0.731)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.098 (0.132)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.34135127067566
==> Statistics for epoch 44: 594 clusters
Epoch: [44][20/200]	Time 0.271 (0.389)	Data 0.001 (0.115)	Loss 0.903 (0.169)
Epoch: [44][40/200]	Time 0.265 (0.371)	Data 0.001 (0.098)	Loss 0.555 (0.464)
Epoch: [44][60/200]	Time 0.269 (0.361)	Data 0.001 (0.089)	Loss 0.862 (0.589)
Epoch: [44][80/200]	Time 0.269 (0.358)	Data 0.001 (0.086)	Loss 0.742 (0.642)
Epoch: [44][100/200]	Time 0.268 (0.354)	Data 0.001 (0.081)	Loss 0.785 (0.655)
Epoch: [44][120/200]	Time 0.265 (0.352)	Data 0.001 (0.080)	Loss 0.709 (0.684)
Epoch: [44][140/200]	Time 0.266 (0.351)	Data 0.000 (0.078)	Loss 0.934 (0.707)
Epoch: [44][160/200]	Time 0.267 (0.351)	Data 0.000 (0.078)	Loss 0.889 (0.715)
Epoch: [44][180/200]	Time 0.269 (0.350)	Data 0.000 (0.078)	Loss 0.977 (0.727)
Epoch: [44][200/200]	Time 0.272 (0.357)	Data 0.001 (0.085)	Loss 0.638 (0.730)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.131 (0.141)	Data 0.035 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.37864661216736
==> Statistics for epoch 45: 594 clusters
Epoch: [45][20/200]	Time 0.272 (0.373)	Data 0.001 (0.102)	Loss 0.695 (0.173)
Epoch: [45][40/200]	Time 0.267 (0.357)	Data 0.001 (0.086)	Loss 0.870 (0.420)
Epoch: [45][60/200]	Time 0.270 (0.355)	Data 0.001 (0.083)	Loss 0.694 (0.556)
Epoch: [45][80/200]	Time 0.268 (0.353)	Data 0.001 (0.081)	Loss 0.871 (0.631)
Epoch: [45][100/200]	Time 0.270 (0.351)	Data 0.001 (0.079)	Loss 0.817 (0.672)
Epoch: [45][120/200]	Time 0.282 (0.351)	Data 0.000 (0.079)	Loss 0.918 (0.695)
Epoch: [45][140/200]	Time 0.268 (0.351)	Data 0.000 (0.078)	Loss 0.946 (0.711)
Epoch: [45][160/200]	Time 0.269 (0.350)	Data 0.000 (0.077)	Loss 0.841 (0.720)
Epoch: [45][180/200]	Time 0.264 (0.350)	Data 0.000 (0.076)	Loss 0.844 (0.727)
Epoch: [45][200/200]	Time 0.282 (0.357)	Data 0.001 (0.083)	Loss 1.209 (0.733)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.098 (0.131)	Data 0.000 (0.032)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.83275032043457
==> Statistics for epoch 46: 595 clusters
Epoch: [46][20/200]	Time 0.278 (0.377)	Data 0.001 (0.103)	Loss 0.661 (0.147)
Epoch: [46][40/200]	Time 0.271 (0.361)	Data 0.002 (0.086)	Loss 0.837 (0.422)
Epoch: [46][60/200]	Time 0.386 (0.355)	Data 0.001 (0.081)	Loss 0.522 (0.540)
Epoch: [46][80/200]	Time 0.275 (0.351)	Data 0.001 (0.077)	Loss 0.643 (0.610)
Epoch: [46][100/200]	Time 0.268 (0.350)	Data 0.001 (0.076)	Loss 0.765 (0.646)
Epoch: [46][120/200]	Time 0.265 (0.350)	Data 0.000 (0.076)	Loss 0.771 (0.679)
Epoch: [46][140/200]	Time 0.262 (0.348)	Data 0.000 (0.075)	Loss 0.581 (0.700)
Epoch: [46][160/200]	Time 0.275 (0.347)	Data 0.000 (0.074)	Loss 0.914 (0.713)
Epoch: [46][180/200]	Time 0.268 (0.347)	Data 0.000 (0.074)	Loss 0.689 (0.718)
Epoch: [46][200/200]	Time 0.282 (0.355)	Data 0.001 (0.081)	Loss 0.674 (0.723)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.142)	Data 0.000 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.775971174240112
==> Statistics for epoch 47: 593 clusters
Epoch: [47][20/200]	Time 0.275 (0.388)	Data 0.001 (0.110)	Loss 0.545 (0.156)
Epoch: [47][40/200]	Time 0.277 (0.368)	Data 0.001 (0.091)	Loss 0.986 (0.453)
Epoch: [47][60/200]	Time 0.376 (0.361)	Data 0.001 (0.085)	Loss 0.662 (0.545)
Epoch: [47][80/200]	Time 0.271 (0.357)	Data 0.001 (0.083)	Loss 0.863 (0.625)
Epoch: [47][100/200]	Time 0.269 (0.355)	Data 0.001 (0.080)	Loss 0.846 (0.659)
Epoch: [47][120/200]	Time 0.266 (0.354)	Data 0.000 (0.079)	Loss 0.738 (0.675)
Epoch: [47][140/200]	Time 0.269 (0.353)	Data 0.000 (0.078)	Loss 0.898 (0.695)
Epoch: [47][160/200]	Time 0.271 (0.352)	Data 0.000 (0.078)	Loss 0.828 (0.713)
Epoch: [47][180/200]	Time 0.266 (0.352)	Data 0.000 (0.078)	Loss 0.630 (0.727)
Epoch: [47][200/200]	Time 0.276 (0.359)	Data 0.001 (0.084)	Loss 1.120 (0.732)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.141)	Data 0.000 (0.041)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.516931772232056
==> Statistics for epoch 48: 592 clusters
Epoch: [48][20/200]	Time 0.271 (0.376)	Data 0.003 (0.106)	Loss 1.159 (0.193)
Epoch: [48][40/200]	Time 0.269 (0.363)	Data 0.001 (0.090)	Loss 0.859 (0.465)
Epoch: [48][60/200]	Time 0.264 (0.356)	Data 0.001 (0.083)	Loss 0.947 (0.586)
Epoch: [48][80/200]	Time 0.268 (0.351)	Data 0.001 (0.079)	Loss 0.687 (0.626)
Epoch: [48][100/200]	Time 0.269 (0.351)	Data 0.001 (0.078)	Loss 1.008 (0.663)
Epoch: [48][120/200]	Time 0.268 (0.350)	Data 0.000 (0.077)	Loss 0.610 (0.685)
Epoch: [48][140/200]	Time 0.267 (0.350)	Data 0.000 (0.077)	Loss 0.786 (0.695)
Epoch: [48][160/200]	Time 0.267 (0.349)	Data 0.000 (0.076)	Loss 0.793 (0.701)
Epoch: [48][180/200]	Time 0.270 (0.349)	Data 0.000 (0.076)	Loss 0.663 (0.711)
Epoch: [48][200/200]	Time 0.267 (0.357)	Data 0.001 (0.083)	Loss 1.101 (0.720)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.097 (0.137)	Data 0.000 (0.036)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.992342472076416
==> Statistics for epoch 49: 593 clusters
Epoch: [49][20/200]	Time 0.271 (0.387)	Data 0.001 (0.109)	Loss 0.762 (0.181)
Epoch: [49][40/200]	Time 0.272 (0.366)	Data 0.001 (0.091)	Loss 0.676 (0.491)
Epoch: [49][60/200]	Time 0.264 (0.358)	Data 0.001 (0.083)	Loss 1.098 (0.602)
Epoch: [49][80/200]	Time 0.266 (0.355)	Data 0.001 (0.081)	Loss 0.785 (0.659)
Epoch: [49][100/200]	Time 0.269 (0.353)	Data 0.001 (0.080)	Loss 0.740 (0.688)
Epoch: [49][120/200]	Time 0.268 (0.353)	Data 0.000 (0.079)	Loss 0.598 (0.696)
Epoch: [49][140/200]	Time 0.270 (0.351)	Data 0.000 (0.078)	Loss 0.666 (0.704)
Epoch: [49][160/200]	Time 0.266 (0.351)	Data 0.000 (0.078)	Loss 0.969 (0.715)
Epoch: [49][180/200]	Time 0.267 (0.350)	Data 0.000 (0.077)	Loss 0.666 (0.726)
Epoch: [49][200/200]	Time 0.274 (0.357)	Data 0.001 (0.084)	Loss 0.967 (0.731)
Extract Features: [50/76]	Time 0.097 (0.140)	Data 0.001 (0.039)	
Mean AP: 90.4%

 * Finished epoch  49  model mAP: 90.4%  best: 90.4% *

==> Test with the best model:
=> Loaded checkpoint 'log/msmt2market/vit_small_cion/model_best.pth.tar'
Extract Features: [50/76]	Time 0.097 (0.144)	Data 0.000 (0.042)	
Mean AP: 90.4%
CMC Scores:
  top-1          95.6%
  top-5          98.0%
  top-10         98.6%
Total running time:  1:25:38.763751
