==========
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_tiny', pretrained_path='/home/ma-user/work/Projects/ReIDNet_Finetune/TransReID/log/bs_exp/ViT_Tiny_MSMT17/64bs_lr0.0004_ep120_warm20_seed0/vit_tiny_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_tiny_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.
ViT Tiny Created!
optimizer: SGD
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.068 (0.438)	Data 0.000 (0.034)	
Computing jaccard distance...
Jaccard distance computing time cost: 21.409247159957886
Clustering criterion: eps: 0.600
==> Statistics for epoch 0: 445 clusters
Epoch: [0][20/200]	Time 0.170 (0.641)	Data 0.000 (0.106)	Loss 7.042 (6.821)
Epoch: [0][40/200]	Time 1.612 (0.481)	Data 1.416 (0.120)	Loss 5.166 (6.516)
Epoch: [0][60/200]	Time 0.177 (0.402)	Data 0.000 (0.100)	Loss 4.589 (5.976)
Epoch: [0][80/200]	Time 0.187 (0.383)	Data 0.001 (0.110)	Loss 4.080 (5.552)
Epoch: [0][100/200]	Time 0.178 (0.355)	Data 0.000 (0.101)	Loss 4.313 (5.268)
Epoch: [0][120/200]	Time 0.197 (0.350)	Data 0.002 (0.107)	Loss 3.544 (5.005)
Epoch: [0][140/200]	Time 0.175 (0.336)	Data 0.000 (0.102)	Loss 4.086 (4.827)
Epoch: [0][160/200]	Time 0.176 (0.332)	Data 0.000 (0.104)	Loss 3.809 (4.658)
Epoch: [0][180/200]	Time 0.167 (0.323)	Data 0.000 (0.100)	Loss 3.836 (4.503)
Epoch: [0][200/200]	Time 0.174 (0.322)	Data 0.000 (0.103)	Loss 3.469 (4.364)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.077 (0.139)	Data 0.000 (0.046)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.650593280792236
==> Statistics for epoch 1: 511 clusters
Epoch: [1][20/200]	Time 0.171 (0.288)	Data 0.001 (0.107)	Loss 4.075 (1.463)
Epoch: [1][40/200]	Time 0.170 (0.272)	Data 0.000 (0.089)	Loss 3.582 (2.282)
Epoch: [1][60/200]	Time 0.175 (0.266)	Data 0.000 (0.082)	Loss 2.994 (2.593)
Epoch: [1][80/200]	Time 0.178 (0.279)	Data 0.001 (0.094)	Loss 3.148 (2.729)
Epoch: [1][100/200]	Time 0.179 (0.272)	Data 0.000 (0.089)	Loss 3.074 (2.781)
Epoch: [1][120/200]	Time 0.178 (0.268)	Data 0.000 (0.085)	Loss 3.438 (2.823)
Epoch: [1][140/200]	Time 0.183 (0.276)	Data 0.001 (0.093)	Loss 2.690 (2.828)
Epoch: [1][160/200]	Time 0.176 (0.273)	Data 0.000 (0.090)	Loss 2.976 (2.833)
Epoch: [1][180/200]	Time 0.295 (0.271)	Data 0.000 (0.088)	Loss 2.816 (2.830)
Epoch: [1][200/200]	Time 0.173 (0.276)	Data 0.000 (0.093)	Loss 3.334 (2.826)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.075 (0.134)	Data 0.000 (0.048)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.576545238494873
==> Statistics for epoch 2: 500 clusters
Epoch: [2][20/200]	Time 0.171 (0.294)	Data 0.000 (0.103)	Loss 3.080 (1.207)
Epoch: [2][40/200]	Time 0.187 (0.275)	Data 0.000 (0.085)	Loss 2.468 (1.925)
Epoch: [2][60/200]	Time 0.170 (0.267)	Data 0.000 (0.079)	Loss 2.910 (2.195)
Epoch: [2][80/200]	Time 0.172 (0.278)	Data 0.000 (0.092)	Loss 2.741 (2.331)
Epoch: [2][100/200]	Time 0.180 (0.272)	Data 0.000 (0.087)	Loss 2.188 (2.382)
Epoch: [2][120/200]	Time 0.171 (0.266)	Data 0.000 (0.082)	Loss 2.491 (2.410)
Epoch: [2][140/200]	Time 0.175 (0.274)	Data 0.000 (0.090)	Loss 2.141 (2.415)
Epoch: [2][160/200]	Time 0.178 (0.272)	Data 0.000 (0.087)	Loss 2.392 (2.438)
Epoch: [2][180/200]	Time 0.295 (0.270)	Data 0.000 (0.085)	Loss 2.468 (2.464)
Epoch: [2][200/200]	Time 0.170 (0.274)	Data 0.000 (0.089)	Loss 2.009 (2.466)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.145 (0.137)	Data 0.070 (0.054)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.566571712493896
==> Statistics for epoch 3: 510 clusters
Epoch: [3][20/200]	Time 0.173 (0.293)	Data 0.001 (0.104)	Loss 2.215 (0.976)
Epoch: [3][40/200]	Time 0.180 (0.274)	Data 0.000 (0.085)	Loss 2.639 (1.682)
Epoch: [3][60/200]	Time 0.172 (0.268)	Data 0.000 (0.081)	Loss 2.676 (1.938)
Epoch: [3][80/200]	Time 0.174 (0.279)	Data 0.001 (0.094)	Loss 2.542 (2.081)
Epoch: [3][100/200]	Time 0.180 (0.274)	Data 0.000 (0.088)	Loss 2.908 (2.154)
Epoch: [3][120/200]	Time 0.170 (0.270)	Data 0.000 (0.086)	Loss 3.147 (2.207)
Epoch: [3][140/200]	Time 0.176 (0.277)	Data 0.001 (0.092)	Loss 2.603 (2.226)
Epoch: [3][160/200]	Time 0.166 (0.273)	Data 0.000 (0.089)	Loss 2.107 (2.235)
Epoch: [3][180/200]	Time 0.303 (0.271)	Data 0.000 (0.087)	Loss 2.032 (2.236)
Epoch: [3][200/200]	Time 0.181 (0.276)	Data 0.001 (0.092)	Loss 2.532 (2.239)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.130)	Data 0.000 (0.048)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.426196098327637
==> Statistics for epoch 4: 521 clusters
Epoch: [4][20/200]	Time 0.178 (0.293)	Data 0.001 (0.104)	Loss 2.414 (0.848)
Epoch: [4][40/200]	Time 0.174 (0.273)	Data 0.001 (0.086)	Loss 1.818 (1.455)
Epoch: [4][60/200]	Time 0.166 (0.267)	Data 0.000 (0.080)	Loss 2.435 (1.739)
Epoch: [4][80/200]	Time 0.185 (0.260)	Data 0.000 (0.076)	Loss 2.521 (1.857)
Epoch: [4][100/200]	Time 0.173 (0.273)	Data 0.001 (0.087)	Loss 2.747 (1.936)
Epoch: [4][120/200]	Time 0.176 (0.269)	Data 0.001 (0.084)	Loss 2.180 (1.970)
Epoch: [4][140/200]	Time 0.168 (0.266)	Data 0.000 (0.081)	Loss 1.753 (2.004)
Epoch: [4][160/200]	Time 0.169 (0.264)	Data 0.000 (0.079)	Loss 2.065 (2.019)
Epoch: [4][180/200]	Time 0.331 (0.270)	Data 0.000 (0.085)	Loss 1.877 (2.035)
Epoch: [4][200/200]	Time 0.170 (0.267)	Data 0.000 (0.083)	Loss 2.318 (2.049)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.069 (0.136)	Data 0.000 (0.053)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.595009326934814
==> Statistics for epoch 5: 547 clusters
Epoch: [5][20/200]	Time 0.191 (0.285)	Data 0.001 (0.097)	Loss 2.221 (0.615)
Epoch: [5][40/200]	Time 0.198 (0.259)	Data 0.001 (0.077)	Loss 2.010 (1.340)
Epoch: [5][60/200]	Time 0.178 (0.255)	Data 0.001 (0.073)	Loss 2.234 (1.584)
Epoch: [5][80/200]	Time 0.175 (0.252)	Data 0.000 (0.071)	Loss 2.284 (1.729)
Epoch: [5][100/200]	Time 0.168 (0.249)	Data 0.000 (0.069)	Loss 2.230 (1.809)
Epoch: [5][120/200]	Time 1.434 (0.259)	Data 1.191 (0.078)	Loss 2.295 (1.849)
Epoch: [5][140/200]	Time 0.172 (0.256)	Data 0.001 (0.076)	Loss 2.384 (1.875)
Epoch: [5][160/200]	Time 0.173 (0.255)	Data 0.001 (0.074)	Loss 2.019 (1.893)
Epoch: [5][180/200]	Time 0.180 (0.254)	Data 0.000 (0.073)	Loss 2.072 (1.923)
Epoch: [5][200/200]	Time 0.173 (0.253)	Data 0.000 (0.072)	Loss 2.803 (1.946)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.200 (0.131)	Data 0.123 (0.051)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.96557092666626
==> Statistics for epoch 6: 548 clusters
Epoch: [6][20/200]	Time 0.196 (0.299)	Data 0.001 (0.107)	Loss 2.019 (0.585)
Epoch: [6][40/200]	Time 0.188 (0.272)	Data 0.001 (0.082)	Loss 1.908 (1.264)
Epoch: [6][60/200]	Time 0.182 (0.263)	Data 0.001 (0.077)	Loss 2.166 (1.508)
Epoch: [6][80/200]	Time 0.179 (0.259)	Data 0.000 (0.073)	Loss 1.787 (1.639)
Epoch: [6][100/200]	Time 0.175 (0.254)	Data 0.000 (0.070)	Loss 2.128 (1.710)
Epoch: [6][120/200]	Time 1.420 (0.263)	Data 1.201 (0.078)	Loss 1.771 (1.769)
Epoch: [6][140/200]	Time 0.173 (0.260)	Data 0.001 (0.075)	Loss 1.813 (1.798)
Epoch: [6][160/200]	Time 0.171 (0.258)	Data 0.000 (0.073)	Loss 2.182 (1.823)
Epoch: [6][180/200]	Time 0.183 (0.256)	Data 0.000 (0.072)	Loss 1.961 (1.835)
Epoch: [6][200/200]	Time 0.171 (0.255)	Data 0.000 (0.071)	Loss 1.971 (1.845)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.071 (0.136)	Data 0.000 (0.055)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.48680090904236
==> Statistics for epoch 7: 547 clusters
Epoch: [7][20/200]	Time 0.208 (0.281)	Data 0.001 (0.100)	Loss 2.209 (0.612)
Epoch: [7][40/200]	Time 0.168 (0.261)	Data 0.000 (0.078)	Loss 1.835 (1.249)
Epoch: [7][60/200]	Time 0.177 (0.257)	Data 0.001 (0.074)	Loss 1.548 (1.501)
Epoch: [7][80/200]	Time 0.180 (0.253)	Data 0.000 (0.070)	Loss 1.930 (1.605)
Epoch: [7][100/200]	Time 0.174 (0.251)	Data 0.000 (0.069)	Loss 1.463 (1.649)
Epoch: [7][120/200]	Time 1.438 (0.261)	Data 1.228 (0.077)	Loss 1.702 (1.698)
Epoch: [7][140/200]	Time 0.175 (0.260)	Data 0.001 (0.076)	Loss 1.681 (1.728)
Epoch: [7][160/200]	Time 0.175 (0.257)	Data 0.001 (0.074)	Loss 1.634 (1.750)
Epoch: [7][180/200]	Time 0.176 (0.256)	Data 0.000 (0.072)	Loss 1.733 (1.770)
Epoch: [7][200/200]	Time 0.174 (0.254)	Data 0.000 (0.072)	Loss 1.938 (1.772)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.159 (0.133)	Data 0.087 (0.050)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.55937933921814
==> Statistics for epoch 8: 551 clusters
Epoch: [8][20/200]	Time 0.186 (0.283)	Data 0.001 (0.097)	Loss 1.654 (0.545)
Epoch: [8][40/200]	Time 0.190 (0.261)	Data 0.001 (0.080)	Loss 1.997 (1.146)
Epoch: [8][60/200]	Time 0.175 (0.254)	Data 0.001 (0.072)	Loss 1.966 (1.376)
Epoch: [8][80/200]	Time 0.299 (0.251)	Data 0.000 (0.068)	Loss 1.660 (1.479)
Epoch: [8][100/200]	Time 0.173 (0.248)	Data 0.000 (0.066)	Loss 1.850 (1.537)
Epoch: [8][120/200]	Time 1.449 (0.258)	Data 1.239 (0.075)	Loss 1.983 (1.588)
Epoch: [8][140/200]	Time 0.171 (0.257)	Data 0.001 (0.073)	Loss 1.976 (1.622)
Epoch: [8][160/200]	Time 0.180 (0.255)	Data 0.001 (0.072)	Loss 1.950 (1.655)
Epoch: [8][180/200]	Time 0.178 (0.254)	Data 0.000 (0.071)	Loss 2.005 (1.664)
Epoch: [8][200/200]	Time 0.172 (0.253)	Data 0.000 (0.070)	Loss 1.324 (1.670)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.073 (0.135)	Data 0.000 (0.053)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.895030736923218
==> Statistics for epoch 9: 553 clusters
Epoch: [9][20/200]	Time 0.178 (0.282)	Data 0.001 (0.101)	Loss 1.446 (0.494)
Epoch: [9][40/200]	Time 0.185 (0.265)	Data 0.001 (0.081)	Loss 1.944 (1.111)
Epoch: [9][60/200]	Time 0.180 (0.259)	Data 0.000 (0.074)	Loss 1.743 (1.329)
Epoch: [9][80/200]	Time 0.173 (0.255)	Data 0.000 (0.071)	Loss 2.097 (1.459)
Epoch: [9][100/200]	Time 0.171 (0.253)	Data 0.000 (0.069)	Loss 1.594 (1.499)
Epoch: [9][120/200]	Time 1.391 (0.262)	Data 1.136 (0.077)	Loss 1.579 (1.545)
Epoch: [9][140/200]	Time 0.172 (0.260)	Data 0.001 (0.075)	Loss 2.285 (1.573)
Epoch: [9][160/200]	Time 0.194 (0.259)	Data 0.001 (0.073)	Loss 1.626 (1.596)
Epoch: [9][180/200]	Time 0.312 (0.258)	Data 0.000 (0.072)	Loss 1.141 (1.613)
Epoch: [9][200/200]	Time 0.168 (0.255)	Data 0.000 (0.071)	Loss 1.251 (1.616)
Extract Features: [50/76]	Time 0.092 (0.135)	Data 0.000 (0.047)	
Mean AP: 85.3%

 * Finished epoch   9  model mAP: 85.3%  best: 85.3% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.132)	Data 0.007 (0.045)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.212066888809204
==> Statistics for epoch 10: 559 clusters
Epoch: [10][20/200]	Time 0.196 (0.325)	Data 0.001 (0.100)	Loss 1.895 (0.520)
Epoch: [10][40/200]	Time 0.324 (0.289)	Data 0.001 (0.081)	Loss 1.704 (1.103)
Epoch: [10][60/200]	Time 0.180 (0.272)	Data 0.001 (0.074)	Loss 2.159 (1.313)
Epoch: [10][80/200]	Time 0.175 (0.266)	Data 0.001 (0.070)	Loss 2.056 (1.412)
Epoch: [10][100/200]	Time 0.175 (0.260)	Data 0.000 (0.068)	Loss 2.026 (1.489)
Epoch: [10][120/200]	Time 1.457 (0.269)	Data 1.243 (0.077)	Loss 1.692 (1.519)
Epoch: [10][140/200]	Time 0.172 (0.265)	Data 0.001 (0.075)	Loss 1.234 (1.517)
Epoch: [10][160/200]	Time 0.174 (0.262)	Data 0.001 (0.073)	Loss 2.234 (1.554)
Epoch: [10][180/200]	Time 0.177 (0.260)	Data 0.001 (0.072)	Loss 1.680 (1.560)
Epoch: [10][200/200]	Time 0.311 (0.259)	Data 0.000 (0.071)	Loss 1.538 (1.575)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.077 (0.139)	Data 0.000 (0.055)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.463284730911255
==> Statistics for epoch 11: 558 clusters
Epoch: [11][20/200]	Time 0.171 (0.288)	Data 0.001 (0.098)	Loss 1.417 (0.500)
Epoch: [11][40/200]	Time 0.175 (0.267)	Data 0.001 (0.082)	Loss 1.688 (1.038)
Epoch: [11][60/200]	Time 0.174 (0.265)	Data 0.001 (0.079)	Loss 1.516 (1.240)
Epoch: [11][80/200]	Time 0.171 (0.260)	Data 0.000 (0.074)	Loss 1.145 (1.353)
Epoch: [11][100/200]	Time 0.175 (0.256)	Data 0.000 (0.071)	Loss 1.288 (1.407)
Epoch: [11][120/200]	Time 1.591 (0.267)	Data 1.378 (0.082)	Loss 1.903 (1.461)
Epoch: [11][140/200]	Time 0.170 (0.263)	Data 0.001 (0.079)	Loss 1.429 (1.480)
Epoch: [11][160/200]	Time 0.187 (0.261)	Data 0.001 (0.077)	Loss 1.618 (1.503)
Epoch: [11][180/200]	Time 0.186 (0.259)	Data 0.001 (0.076)	Loss 1.773 (1.527)
Epoch: [11][200/200]	Time 0.176 (0.258)	Data 0.000 (0.075)	Loss 1.677 (1.540)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.135)	Data 0.000 (0.047)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.600944995880127
==> Statistics for epoch 12: 560 clusters
Epoch: [12][20/200]	Time 0.175 (0.281)	Data 0.001 (0.094)	Loss 1.527 (0.442)
Epoch: [12][40/200]	Time 0.174 (0.262)	Data 0.001 (0.081)	Loss 1.211 (0.980)
Epoch: [12][60/200]	Time 0.180 (0.256)	Data 0.001 (0.074)	Loss 1.599 (1.184)
Epoch: [12][80/200]	Time 0.179 (0.255)	Data 0.000 (0.073)	Loss 1.607 (1.272)
Epoch: [12][100/200]	Time 0.173 (0.253)	Data 0.000 (0.071)	Loss 1.692 (1.330)
Epoch: [12][120/200]	Time 1.492 (0.263)	Data 1.290 (0.080)	Loss 1.598 (1.378)
Epoch: [12][140/200]	Time 0.171 (0.260)	Data 0.001 (0.078)	Loss 1.210 (1.410)
Epoch: [12][160/200]	Time 0.179 (0.258)	Data 0.001 (0.076)	Loss 1.512 (1.439)
Epoch: [12][180/200]	Time 0.171 (0.257)	Data 0.000 (0.075)	Loss 1.543 (1.450)
Epoch: [12][200/200]	Time 0.175 (0.256)	Data 0.000 (0.074)	Loss 1.365 (1.454)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.139)	Data 0.000 (0.060)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.819933652877808
==> Statistics for epoch 13: 556 clusters
Epoch: [13][20/200]	Time 0.180 (0.277)	Data 0.000 (0.095)	Loss 1.720 (0.419)
Epoch: [13][40/200]	Time 0.182 (0.261)	Data 0.001 (0.077)	Loss 1.387 (0.930)
Epoch: [13][60/200]	Time 0.176 (0.256)	Data 0.001 (0.071)	Loss 1.535 (1.122)
Epoch: [13][80/200]	Time 0.179 (0.252)	Data 0.000 (0.069)	Loss 1.652 (1.216)
Epoch: [13][100/200]	Time 0.173 (0.250)	Data 0.000 (0.067)	Loss 1.635 (1.278)
Epoch: [13][120/200]	Time 1.494 (0.261)	Data 1.287 (0.078)	Loss 1.351 (1.325)
Epoch: [13][140/200]	Time 0.173 (0.258)	Data 0.001 (0.076)	Loss 1.600 (1.352)
Epoch: [13][160/200]	Time 0.178 (0.257)	Data 0.000 (0.074)	Loss 1.358 (1.375)
Epoch: [13][180/200]	Time 0.175 (0.256)	Data 0.000 (0.073)	Loss 1.398 (1.385)
Epoch: [13][200/200]	Time 0.281 (0.255)	Data 0.000 (0.071)	Loss 1.573 (1.398)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.069 (0.136)	Data 0.000 (0.053)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.978259563446045
==> Statistics for epoch 14: 554 clusters
Epoch: [14][20/200]	Time 0.189 (0.294)	Data 0.001 (0.102)	Loss 1.378 (0.408)
Epoch: [14][40/200]	Time 0.176 (0.270)	Data 0.001 (0.084)	Loss 1.413 (0.892)
Epoch: [14][60/200]	Time 0.173 (0.265)	Data 0.001 (0.078)	Loss 1.388 (1.099)
Epoch: [14][80/200]	Time 0.170 (0.261)	Data 0.000 (0.074)	Loss 1.051 (1.221)
Epoch: [14][100/200]	Time 0.172 (0.258)	Data 0.000 (0.074)	Loss 1.899 (1.278)
Epoch: [14][120/200]	Time 1.345 (0.266)	Data 1.104 (0.080)	Loss 1.127 (1.305)
Epoch: [14][140/200]	Time 0.183 (0.262)	Data 0.001 (0.077)	Loss 1.803 (1.333)
Epoch: [14][160/200]	Time 0.183 (0.261)	Data 0.001 (0.076)	Loss 1.325 (1.347)
Epoch: [14][180/200]	Time 0.172 (0.258)	Data 0.000 (0.074)	Loss 2.091 (1.370)
Epoch: [14][200/200]	Time 0.172 (0.257)	Data 0.000 (0.073)	Loss 1.467 (1.379)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.068 (0.133)	Data 0.000 (0.051)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.52122139930725
==> Statistics for epoch 15: 555 clusters
Epoch: [15][20/200]	Time 0.183 (0.283)	Data 0.001 (0.093)	Loss 1.469 (0.420)
Epoch: [15][40/200]	Time 0.181 (0.261)	Data 0.001 (0.075)	Loss 1.689 (0.962)
Epoch: [15][60/200]	Time 0.320 (0.256)	Data 0.001 (0.070)	Loss 1.632 (1.155)
Epoch: [15][80/200]	Time 0.194 (0.253)	Data 0.000 (0.069)	Loss 1.516 (1.211)
Epoch: [15][100/200]	Time 0.181 (0.252)	Data 0.000 (0.068)	Loss 1.681 (1.258)
Epoch: [15][120/200]	Time 1.526 (0.262)	Data 1.304 (0.078)	Loss 1.165 (1.277)
Epoch: [15][140/200]	Time 0.193 (0.261)	Data 0.001 (0.077)	Loss 1.742 (1.318)
Epoch: [15][160/200]	Time 0.186 (0.258)	Data 0.002 (0.074)	Loss 1.137 (1.337)
Epoch: [15][180/200]	Time 0.184 (0.257)	Data 0.000 (0.073)	Loss 1.309 (1.345)
Epoch: [15][200/200]	Time 0.170 (0.256)	Data 0.000 (0.072)	Loss 1.048 (1.348)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.108 (0.135)	Data 0.032 (0.051)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.835131883621216
==> Statistics for epoch 16: 566 clusters
Epoch: [16][20/200]	Time 0.178 (0.286)	Data 0.001 (0.100)	Loss 1.685 (0.410)
Epoch: [16][40/200]	Time 0.204 (0.268)	Data 0.001 (0.081)	Loss 1.421 (0.869)
Epoch: [16][60/200]	Time 0.175 (0.258)	Data 0.001 (0.074)	Loss 1.454 (1.061)
Epoch: [16][80/200]	Time 0.181 (0.258)	Data 0.000 (0.072)	Loss 1.757 (1.154)
Epoch: [16][100/200]	Time 0.169 (0.255)	Data 0.000 (0.069)	Loss 1.568 (1.217)
Epoch: [16][120/200]	Time 1.380 (0.263)	Data 1.148 (0.078)	Loss 1.394 (1.245)
Epoch: [16][140/200]	Time 0.178 (0.261)	Data 0.000 (0.076)	Loss 1.136 (1.269)
Epoch: [16][160/200]	Time 0.170 (0.260)	Data 0.001 (0.074)	Loss 1.338 (1.292)
Epoch: [16][180/200]	Time 0.183 (0.258)	Data 0.001 (0.073)	Loss 1.443 (1.299)
Epoch: [16][200/200]	Time 0.168 (0.256)	Data 0.000 (0.072)	Loss 1.303 (1.310)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.074 (0.135)	Data 0.000 (0.056)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.999473810195923
==> Statistics for epoch 17: 566 clusters
Epoch: [17][20/200]	Time 0.187 (0.283)	Data 0.001 (0.100)	Loss 1.253 (0.364)
Epoch: [17][40/200]	Time 0.185 (0.267)	Data 0.001 (0.081)	Loss 1.186 (0.843)
Epoch: [17][60/200]	Time 0.176 (0.256)	Data 0.001 (0.073)	Loss 1.654 (1.014)
Epoch: [17][80/200]	Time 0.169 (0.252)	Data 0.000 (0.070)	Loss 1.601 (1.089)
Epoch: [17][100/200]	Time 0.173 (0.251)	Data 0.000 (0.068)	Loss 1.497 (1.157)
Epoch: [17][120/200]	Time 1.383 (0.259)	Data 1.157 (0.077)	Loss 1.767 (1.205)
Epoch: [17][140/200]	Time 0.177 (0.257)	Data 0.001 (0.075)	Loss 1.882 (1.232)
Epoch: [17][160/200]	Time 0.177 (0.256)	Data 0.001 (0.073)	Loss 1.357 (1.246)
Epoch: [17][180/200]	Time 0.179 (0.254)	Data 0.001 (0.072)	Loss 1.384 (1.267)
Epoch: [17][200/200]	Time 0.174 (0.254)	Data 0.000 (0.071)	Loss 1.151 (1.278)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.133)	Data 0.000 (0.050)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.942012310028076
==> Statistics for epoch 18: 562 clusters
Epoch: [18][20/200]	Time 0.178 (0.274)	Data 0.001 (0.093)	Loss 1.611 (0.396)
Epoch: [18][40/200]	Time 0.181 (0.260)	Data 0.001 (0.077)	Loss 1.251 (0.852)
Epoch: [18][60/200]	Time 0.173 (0.257)	Data 0.001 (0.073)	Loss 1.188 (1.030)
Epoch: [18][80/200]	Time 0.178 (0.251)	Data 0.000 (0.069)	Loss 1.214 (1.117)
Epoch: [18][100/200]	Time 0.172 (0.251)	Data 0.000 (0.068)	Loss 1.525 (1.166)
Epoch: [18][120/200]	Time 1.410 (0.261)	Data 1.191 (0.077)	Loss 1.450 (1.188)
Epoch: [18][140/200]	Time 0.175 (0.258)	Data 0.001 (0.075)	Loss 1.194 (1.210)
Epoch: [18][160/200]	Time 0.189 (0.256)	Data 0.001 (0.073)	Loss 1.220 (1.235)
Epoch: [18][180/200]	Time 0.168 (0.255)	Data 0.000 (0.072)	Loss 1.625 (1.256)
Epoch: [18][200/200]	Time 0.165 (0.254)	Data 0.000 (0.071)	Loss 1.336 (1.264)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.134)	Data 0.000 (0.052)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.320533990859985
==> Statistics for epoch 19: 563 clusters
Epoch: [19][20/200]	Time 0.182 (0.274)	Data 0.000 (0.095)	Loss 1.770 (0.389)
Epoch: [19][40/200]	Time 0.185 (0.262)	Data 0.001 (0.077)	Loss 1.369 (0.863)
Epoch: [19][60/200]	Time 0.183 (0.257)	Data 0.001 (0.072)	Loss 1.248 (1.033)
Epoch: [19][80/200]	Time 0.169 (0.254)	Data 0.000 (0.069)	Loss 1.333 (1.103)
Epoch: [19][100/200]	Time 0.175 (0.251)	Data 0.000 (0.067)	Loss 1.431 (1.142)
Epoch: [19][120/200]	Time 1.490 (0.262)	Data 1.274 (0.078)	Loss 1.137 (1.166)
Epoch: [19][140/200]	Time 0.177 (0.261)	Data 0.000 (0.076)	Loss 1.517 (1.196)
Epoch: [19][160/200]	Time 0.171 (0.259)	Data 0.000 (0.075)	Loss 1.616 (1.222)
Epoch: [19][180/200]	Time 0.180 (0.258)	Data 0.000 (0.074)	Loss 1.524 (1.233)
Epoch: [19][200/200]	Time 0.165 (0.256)	Data 0.000 (0.073)	Loss 1.498 (1.248)
Extract Features: [50/76]	Time 0.076 (0.138)	Data 0.000 (0.056)	
Mean AP: 87.0%

 * Finished epoch  19  model mAP: 87.0%  best: 87.0% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.072 (0.131)	Data 0.000 (0.050)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.35539484024048
==> Statistics for epoch 20: 569 clusters
Epoch: [20][20/200]	Time 0.185 (0.285)	Data 0.000 (0.102)	Loss 1.160 (0.354)
Epoch: [20][40/200]	Time 0.168 (0.265)	Data 0.000 (0.083)	Loss 1.531 (0.831)
Epoch: [20][60/200]	Time 0.176 (0.257)	Data 0.000 (0.076)	Loss 1.423 (0.957)
Epoch: [20][80/200]	Time 0.175 (0.255)	Data 0.000 (0.073)	Loss 1.213 (1.027)
Epoch: [20][100/200]	Time 0.177 (0.253)	Data 0.000 (0.072)	Loss 1.138 (1.087)
Epoch: [20][120/200]	Time 1.633 (0.264)	Data 1.428 (0.083)	Loss 1.141 (1.103)
Epoch: [20][140/200]	Time 0.184 (0.262)	Data 0.001 (0.080)	Loss 1.265 (1.129)
Epoch: [20][160/200]	Time 0.167 (0.261)	Data 0.001 (0.079)	Loss 1.355 (1.146)
Epoch: [20][180/200]	Time 0.169 (0.259)	Data 0.000 (0.078)	Loss 1.753 (1.164)
Epoch: [20][200/200]	Time 0.176 (0.259)	Data 0.000 (0.077)	Loss 1.507 (1.170)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.136)	Data 0.000 (0.057)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.975475072860718
==> Statistics for epoch 21: 573 clusters
Epoch: [21][20/200]	Time 0.173 (0.275)	Data 0.001 (0.095)	Loss 1.184 (0.323)
Epoch: [21][40/200]	Time 0.185 (0.264)	Data 0.001 (0.082)	Loss 1.745 (0.761)
Epoch: [21][60/200]	Time 0.185 (0.259)	Data 0.001 (0.076)	Loss 1.526 (0.931)
Epoch: [21][80/200]	Time 0.187 (0.258)	Data 0.000 (0.075)	Loss 0.986 (0.987)
Epoch: [21][100/200]	Time 0.168 (0.255)	Data 0.000 (0.073)	Loss 1.321 (1.037)
Epoch: [21][120/200]	Time 1.563 (0.266)	Data 1.326 (0.083)	Loss 1.153 (1.066)
Epoch: [21][140/200]	Time 0.170 (0.264)	Data 0.001 (0.081)	Loss 1.334 (1.086)
Epoch: [21][160/200]	Time 0.173 (0.263)	Data 0.001 (0.080)	Loss 1.222 (1.101)
Epoch: [21][180/200]	Time 0.176 (0.262)	Data 0.001 (0.078)	Loss 1.254 (1.122)
Epoch: [21][200/200]	Time 0.182 (0.260)	Data 0.000 (0.077)	Loss 1.149 (1.132)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.071 (0.139)	Data 0.000 (0.059)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.886552095413208
==> Statistics for epoch 22: 567 clusters
Epoch: [22][20/200]	Time 0.201 (0.303)	Data 0.001 (0.111)	Loss 1.390 (0.305)
Epoch: [22][40/200]	Time 0.180 (0.277)	Data 0.001 (0.087)	Loss 1.104 (0.742)
Epoch: [22][60/200]	Time 0.177 (0.266)	Data 0.001 (0.079)	Loss 1.455 (0.917)
Epoch: [22][80/200]	Time 0.179 (0.262)	Data 0.000 (0.076)	Loss 1.314 (0.991)
Epoch: [22][100/200]	Time 0.166 (0.260)	Data 0.000 (0.075)	Loss 0.788 (1.015)
Epoch: [22][120/200]	Time 1.606 (0.270)	Data 1.402 (0.085)	Loss 0.619 (1.037)
Epoch: [22][140/200]	Time 0.171 (0.267)	Data 0.001 (0.083)	Loss 0.973 (1.059)
Epoch: [22][160/200]	Time 0.179 (0.266)	Data 0.001 (0.081)	Loss 1.053 (1.069)
Epoch: [22][180/200]	Time 0.181 (0.264)	Data 0.000 (0.080)	Loss 0.946 (1.085)
Epoch: [22][200/200]	Time 0.175 (0.263)	Data 0.000 (0.078)	Loss 1.035 (1.100)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.075 (0.135)	Data 0.000 (0.053)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.989075660705566
==> Statistics for epoch 23: 568 clusters
Epoch: [23][20/200]	Time 0.174 (0.290)	Data 0.001 (0.102)	Loss 1.196 (0.306)
Epoch: [23][40/200]	Time 0.172 (0.271)	Data 0.001 (0.083)	Loss 1.171 (0.757)
Epoch: [23][60/200]	Time 0.195 (0.263)	Data 0.001 (0.078)	Loss 0.964 (0.887)
Epoch: [23][80/200]	Time 0.176 (0.258)	Data 0.000 (0.073)	Loss 0.943 (0.946)
Epoch: [23][100/200]	Time 0.172 (0.253)	Data 0.000 (0.070)	Loss 1.414 (0.999)
Epoch: [23][120/200]	Time 1.592 (0.264)	Data 1.374 (0.081)	Loss 1.088 (1.034)
Epoch: [23][140/200]	Time 0.192 (0.261)	Data 0.001 (0.078)	Loss 1.019 (1.057)
Epoch: [23][160/200]	Time 0.190 (0.260)	Data 0.001 (0.077)	Loss 1.534 (1.081)
Epoch: [23][180/200]	Time 0.181 (0.259)	Data 0.001 (0.076)	Loss 1.411 (1.092)
Epoch: [23][200/200]	Time 0.174 (0.258)	Data 0.000 (0.075)	Loss 1.442 (1.104)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.071 (0.138)	Data 0.000 (0.058)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.986151933670044
==> Statistics for epoch 24: 569 clusters
Epoch: [24][20/200]	Time 0.185 (0.280)	Data 0.000 (0.096)	Loss 1.185 (0.330)
Epoch: [24][40/200]	Time 0.183 (0.269)	Data 0.000 (0.084)	Loss 1.210 (0.750)
Epoch: [24][60/200]	Time 0.180 (0.261)	Data 0.001 (0.077)	Loss 1.521 (0.896)
Epoch: [24][80/200]	Time 0.169 (0.260)	Data 0.000 (0.075)	Loss 1.037 (0.973)
Epoch: [24][100/200]	Time 0.176 (0.258)	Data 0.000 (0.073)	Loss 1.362 (1.026)
Epoch: [24][120/200]	Time 1.504 (0.267)	Data 1.265 (0.082)	Loss 1.209 (1.059)
Epoch: [24][140/200]	Time 0.176 (0.265)	Data 0.001 (0.080)	Loss 0.815 (1.066)
Epoch: [24][160/200]	Time 0.178 (0.264)	Data 0.000 (0.078)	Loss 1.259 (1.078)
Epoch: [24][180/200]	Time 0.177 (0.263)	Data 0.000 (0.077)	Loss 1.601 (1.102)
Epoch: [24][200/200]	Time 0.166 (0.262)	Data 0.000 (0.077)	Loss 1.225 (1.117)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.139)	Data 0.000 (0.056)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.08930015563965
==> Statistics for epoch 25: 573 clusters
Epoch: [25][20/200]	Time 0.197 (0.294)	Data 0.001 (0.100)	Loss 1.096 (0.337)
Epoch: [25][40/200]	Time 0.186 (0.271)	Data 0.001 (0.083)	Loss 1.153 (0.724)
Epoch: [25][60/200]	Time 0.177 (0.266)	Data 0.001 (0.080)	Loss 1.266 (0.857)
Epoch: [25][80/200]	Time 0.176 (0.263)	Data 0.000 (0.076)	Loss 1.163 (0.955)
Epoch: [25][100/200]	Time 0.170 (0.259)	Data 0.000 (0.074)	Loss 1.121 (1.004)
Epoch: [25][120/200]	Time 1.525 (0.269)	Data 1.285 (0.083)	Loss 0.851 (1.032)
Epoch: [25][140/200]	Time 0.177 (0.267)	Data 0.001 (0.081)	Loss 0.787 (1.055)
Epoch: [25][160/200]	Time 0.174 (0.264)	Data 0.001 (0.078)	Loss 1.743 (1.079)
Epoch: [25][180/200]	Time 0.176 (0.263)	Data 0.001 (0.077)	Loss 1.179 (1.101)
Epoch: [25][200/200]	Time 0.173 (0.261)	Data 0.000 (0.076)	Loss 1.147 (1.104)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.089 (0.135)	Data 0.017 (0.049)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.168365716934204
==> Statistics for epoch 26: 570 clusters
Epoch: [26][20/200]	Time 0.196 (0.292)	Data 0.001 (0.099)	Loss 1.423 (0.315)
Epoch: [26][40/200]	Time 0.194 (0.273)	Data 0.001 (0.083)	Loss 1.317 (0.727)
Epoch: [26][60/200]	Time 0.177 (0.265)	Data 0.001 (0.077)	Loss 0.965 (0.859)
Epoch: [26][80/200]	Time 0.180 (0.262)	Data 0.000 (0.074)	Loss 1.183 (0.923)
Epoch: [26][100/200]	Time 0.175 (0.260)	Data 0.000 (0.073)	Loss 1.153 (0.989)
Epoch: [26][120/200]	Time 1.611 (0.268)	Data 1.267 (0.081)	Loss 0.979 (1.032)
Epoch: [26][140/200]	Time 0.170 (0.265)	Data 0.001 (0.079)	Loss 1.182 (1.057)
Epoch: [26][160/200]	Time 0.178 (0.263)	Data 0.001 (0.077)	Loss 1.491 (1.065)
Epoch: [26][180/200]	Time 0.173 (0.262)	Data 0.001 (0.076)	Loss 1.580 (1.079)
Epoch: [26][200/200]	Time 0.174 (0.260)	Data 0.000 (0.075)	Loss 1.331 (1.097)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.102 (0.135)	Data 0.027 (0.053)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.307537078857422
==> Statistics for epoch 27: 567 clusters
Epoch: [27][20/200]	Time 0.194 (0.287)	Data 0.001 (0.102)	Loss 0.790 (0.304)
Epoch: [27][40/200]	Time 0.182 (0.270)	Data 0.001 (0.085)	Loss 1.106 (0.737)
Epoch: [27][60/200]	Time 0.182 (0.262)	Data 0.001 (0.080)	Loss 0.984 (0.916)
Epoch: [27][80/200]	Time 0.181 (0.259)	Data 0.000 (0.076)	Loss 1.427 (0.985)
Epoch: [27][100/200]	Time 0.173 (0.258)	Data 0.000 (0.074)	Loss 1.021 (1.024)
Epoch: [27][120/200]	Time 1.513 (0.267)	Data 1.310 (0.083)	Loss 0.931 (1.040)
Epoch: [27][140/200]	Time 0.179 (0.264)	Data 0.001 (0.081)	Loss 1.116 (1.057)
Epoch: [27][160/200]	Time 0.185 (0.263)	Data 0.000 (0.079)	Loss 1.229 (1.065)
Epoch: [27][180/200]	Time 0.180 (0.261)	Data 0.001 (0.078)	Loss 1.317 (1.079)
Epoch: [27][200/200]	Time 0.175 (0.259)	Data 0.000 (0.076)	Loss 1.206 (1.089)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.214 (0.136)	Data 0.142 (0.049)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.469817399978638
==> Statistics for epoch 28: 569 clusters
Epoch: [28][20/200]	Time 0.181 (0.299)	Data 0.001 (0.110)	Loss 1.409 (0.331)
Epoch: [28][40/200]	Time 0.189 (0.273)	Data 0.001 (0.085)	Loss 0.903 (0.703)
Epoch: [28][60/200]	Time 0.175 (0.264)	Data 0.001 (0.078)	Loss 1.205 (0.826)
Epoch: [28][80/200]	Time 0.175 (0.261)	Data 0.000 (0.076)	Loss 1.210 (0.915)
Epoch: [28][100/200]	Time 0.176 (0.259)	Data 0.000 (0.073)	Loss 1.047 (0.973)
Epoch: [28][120/200]	Time 1.640 (0.270)	Data 1.439 (0.084)	Loss 1.281 (1.021)
Epoch: [28][140/200]	Time 0.182 (0.268)	Data 0.001 (0.083)	Loss 0.983 (1.043)
Epoch: [28][160/200]	Time 0.167 (0.266)	Data 0.001 (0.081)	Loss 1.673 (1.069)
Epoch: [28][180/200]	Time 0.176 (0.263)	Data 0.000 (0.080)	Loss 1.258 (1.085)
Epoch: [28][200/200]	Time 0.166 (0.261)	Data 0.000 (0.078)	Loss 0.849 (1.094)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.090 (0.133)	Data 0.011 (0.049)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.26088285446167
==> Statistics for epoch 29: 571 clusters
Epoch: [29][20/200]	Time 0.191 (0.280)	Data 0.001 (0.096)	Loss 1.129 (0.329)
Epoch: [29][40/200]	Time 0.182 (0.266)	Data 0.001 (0.080)	Loss 1.088 (0.726)
Epoch: [29][60/200]	Time 0.177 (0.261)	Data 0.001 (0.078)	Loss 1.258 (0.850)
Epoch: [29][80/200]	Time 0.169 (0.256)	Data 0.000 (0.073)	Loss 1.453 (0.933)
Epoch: [29][100/200]	Time 0.168 (0.255)	Data 0.000 (0.071)	Loss 1.203 (0.992)
Epoch: [29][120/200]	Time 1.423 (0.263)	Data 1.196 (0.080)	Loss 1.489 (1.028)
Epoch: [29][140/200]	Time 0.178 (0.262)	Data 0.001 (0.078)	Loss 1.367 (1.053)
Epoch: [29][160/200]	Time 0.196 (0.261)	Data 0.001 (0.077)	Loss 1.437 (1.066)
Epoch: [29][180/200]	Time 0.171 (0.260)	Data 0.001 (0.075)	Loss 1.022 (1.069)
Epoch: [29][200/200]	Time 0.173 (0.259)	Data 0.000 (0.074)	Loss 1.184 (1.084)
Extract Features: [50/76]	Time 0.075 (0.133)	Data 0.000 (0.049)	
Mean AP: 87.8%

 * Finished epoch  29  model mAP: 87.8%  best: 87.8% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.137)	Data 0.000 (0.054)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.050004720687866
==> Statistics for epoch 30: 570 clusters
Epoch: [30][20/200]	Time 0.169 (0.274)	Data 0.001 (0.092)	Loss 1.116 (0.311)
Epoch: [30][40/200]	Time 0.175 (0.268)	Data 0.001 (0.082)	Loss 1.610 (0.706)
Epoch: [30][60/200]	Time 0.174 (0.264)	Data 0.001 (0.077)	Loss 1.055 (0.837)
Epoch: [30][80/200]	Time 0.166 (0.257)	Data 0.000 (0.073)	Loss 1.230 (0.923)
Epoch: [30][100/200]	Time 0.173 (0.257)	Data 0.000 (0.072)	Loss 1.327 (0.977)
Epoch: [30][120/200]	Time 1.506 (0.269)	Data 1.273 (0.082)	Loss 0.888 (0.999)
Epoch: [30][140/200]	Time 0.178 (0.265)	Data 0.001 (0.080)	Loss 1.083 (1.030)
Epoch: [30][160/200]	Time 0.185 (0.263)	Data 0.001 (0.078)	Loss 1.283 (1.044)
Epoch: [30][180/200]	Time 0.186 (0.262)	Data 0.001 (0.077)	Loss 1.022 (1.059)
Epoch: [30][200/200]	Time 0.177 (0.261)	Data 0.000 (0.076)	Loss 1.181 (1.072)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.123 (0.137)	Data 0.053 (0.055)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.014378786087036
==> Statistics for epoch 31: 570 clusters
Epoch: [31][20/200]	Time 0.198 (0.283)	Data 0.001 (0.100)	Loss 1.142 (0.310)
Epoch: [31][40/200]	Time 0.180 (0.268)	Data 0.001 (0.083)	Loss 1.181 (0.724)
Epoch: [31][60/200]	Time 0.189 (0.261)	Data 0.001 (0.078)	Loss 0.964 (0.864)
Epoch: [31][80/200]	Time 0.170 (0.260)	Data 0.000 (0.076)	Loss 1.336 (0.951)
Epoch: [31][100/200]	Time 0.173 (0.259)	Data 0.000 (0.075)	Loss 1.279 (1.000)
Epoch: [31][120/200]	Time 1.395 (0.265)	Data 1.159 (0.082)	Loss 1.175 (1.028)
Epoch: [31][140/200]	Time 0.179 (0.262)	Data 0.001 (0.079)	Loss 1.045 (1.046)
Epoch: [31][160/200]	Time 0.177 (0.261)	Data 0.001 (0.077)	Loss 0.936 (1.053)
Epoch: [31][180/200]	Time 0.183 (0.259)	Data 0.001 (0.076)	Loss 1.155 (1.062)
Epoch: [31][200/200]	Time 0.183 (0.259)	Data 0.000 (0.076)	Loss 1.123 (1.076)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.069 (0.135)	Data 0.000 (0.055)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.95554757118225
==> Statistics for epoch 32: 570 clusters
Epoch: [32][20/200]	Time 0.181 (0.283)	Data 0.001 (0.100)	Loss 1.586 (0.314)
Epoch: [32][40/200]	Time 0.186 (0.269)	Data 0.001 (0.084)	Loss 1.518 (0.702)
Epoch: [32][60/200]	Time 0.177 (0.261)	Data 0.001 (0.075)	Loss 1.131 (0.843)
Epoch: [32][80/200]	Time 0.181 (0.258)	Data 0.000 (0.074)	Loss 1.140 (0.919)
Epoch: [32][100/200]	Time 0.174 (0.256)	Data 0.000 (0.071)	Loss 1.523 (0.970)
Epoch: [32][120/200]	Time 1.708 (0.268)	Data 1.502 (0.082)	Loss 0.898 (1.010)
Epoch: [32][140/200]	Time 0.174 (0.267)	Data 0.001 (0.081)	Loss 0.776 (1.035)
Epoch: [32][160/200]	Time 0.192 (0.265)	Data 0.001 (0.079)	Loss 1.411 (1.043)
Epoch: [32][180/200]	Time 0.180 (0.264)	Data 0.001 (0.078)	Loss 0.989 (1.055)
Epoch: [32][200/200]	Time 0.172 (0.263)	Data 0.000 (0.077)	Loss 0.797 (1.064)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.255 (0.139)	Data 0.068 (0.055)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.92829728126526
==> Statistics for epoch 33: 571 clusters
Epoch: [33][20/200]	Time 0.181 (0.284)	Data 0.001 (0.100)	Loss 1.036 (0.321)
Epoch: [33][40/200]	Time 0.176 (0.271)	Data 0.001 (0.085)	Loss 0.847 (0.714)
Epoch: [33][60/200]	Time 0.177 (0.264)	Data 0.001 (0.078)	Loss 1.167 (0.883)
Epoch: [33][80/200]	Time 0.180 (0.259)	Data 0.000 (0.075)	Loss 1.456 (0.946)
Epoch: [33][100/200]	Time 0.167 (0.258)	Data 0.000 (0.073)	Loss 1.328 (0.994)
Epoch: [33][120/200]	Time 1.588 (0.269)	Data 1.374 (0.084)	Loss 1.266 (1.025)
Epoch: [33][140/200]	Time 0.172 (0.266)	Data 0.001 (0.081)	Loss 1.373 (1.045)
Epoch: [33][160/200]	Time 0.186 (0.264)	Data 0.001 (0.079)	Loss 0.850 (1.066)
Epoch: [33][180/200]	Time 0.180 (0.263)	Data 0.001 (0.078)	Loss 1.062 (1.076)
Epoch: [33][200/200]	Time 0.177 (0.261)	Data 0.000 (0.077)	Loss 1.354 (1.082)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.131 (0.137)	Data 0.059 (0.054)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.325875282287598
==> Statistics for epoch 34: 573 clusters
Epoch: [34][20/200]	Time 0.192 (0.297)	Data 0.001 (0.109)	Loss 1.073 (0.309)
Epoch: [34][40/200]	Time 0.187 (0.274)	Data 0.001 (0.086)	Loss 1.092 (0.731)
Epoch: [34][60/200]	Time 0.189 (0.265)	Data 0.001 (0.080)	Loss 1.217 (0.873)
Epoch: [34][80/200]	Time 0.176 (0.262)	Data 0.000 (0.078)	Loss 1.304 (0.973)
Epoch: [34][100/200]	Time 0.170 (0.260)	Data 0.000 (0.076)	Loss 1.118 (1.002)
Epoch: [34][120/200]	Time 1.606 (0.269)	Data 1.402 (0.086)	Loss 1.016 (1.022)
Epoch: [34][140/200]	Time 0.180 (0.267)	Data 0.001 (0.083)	Loss 1.185 (1.044)
Epoch: [34][160/200]	Time 0.179 (0.266)	Data 0.001 (0.082)	Loss 1.267 (1.058)
Epoch: [34][180/200]	Time 0.188 (0.263)	Data 0.001 (0.079)	Loss 1.074 (1.060)
Epoch: [34][200/200]	Time 0.175 (0.261)	Data 0.000 (0.078)	Loss 1.050 (1.065)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.136)	Data 0.000 (0.051)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.015896320343018
==> Statistics for epoch 35: 570 clusters
Epoch: [35][20/200]	Time 0.205 (0.300)	Data 0.001 (0.108)	Loss 1.204 (0.309)
Epoch: [35][40/200]	Time 0.170 (0.277)	Data 0.001 (0.090)	Loss 1.080 (0.721)
Epoch: [35][60/200]	Time 0.168 (0.268)	Data 0.001 (0.082)	Loss 1.335 (0.874)
Epoch: [35][80/200]	Time 0.180 (0.262)	Data 0.000 (0.077)	Loss 0.981 (0.941)
Epoch: [35][100/200]	Time 0.168 (0.257)	Data 0.000 (0.073)	Loss 0.878 (0.983)
Epoch: [35][120/200]	Time 1.576 (0.268)	Data 1.358 (0.083)	Loss 1.100 (0.999)
Epoch: [35][140/200]	Time 0.171 (0.265)	Data 0.001 (0.081)	Loss 1.227 (1.021)
Epoch: [35][160/200]	Time 0.183 (0.264)	Data 0.001 (0.080)	Loss 1.329 (1.046)
Epoch: [35][180/200]	Time 0.177 (0.262)	Data 0.001 (0.077)	Loss 1.172 (1.053)
Epoch: [35][200/200]	Time 0.171 (0.261)	Data 0.000 (0.076)	Loss 0.950 (1.056)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.075 (0.142)	Data 0.000 (0.059)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.235573530197144
==> Statistics for epoch 36: 571 clusters
Epoch: [36][20/200]	Time 0.181 (0.290)	Data 0.001 (0.100)	Loss 0.771 (0.284)
Epoch: [36][40/200]	Time 0.178 (0.272)	Data 0.001 (0.084)	Loss 0.786 (0.681)
Epoch: [36][60/200]	Time 0.325 (0.263)	Data 0.001 (0.076)	Loss 0.994 (0.839)
Epoch: [36][80/200]	Time 0.165 (0.259)	Data 0.000 (0.075)	Loss 1.226 (0.919)
Epoch: [36][100/200]	Time 0.171 (0.258)	Data 0.000 (0.073)	Loss 1.119 (0.955)
Epoch: [36][120/200]	Time 1.614 (0.268)	Data 1.409 (0.083)	Loss 1.111 (0.978)
Epoch: [36][140/200]	Time 0.178 (0.266)	Data 0.001 (0.081)	Loss 1.063 (1.005)
Epoch: [36][160/200]	Time 0.179 (0.264)	Data 0.001 (0.080)	Loss 0.959 (1.017)
Epoch: [36][180/200]	Time 0.178 (0.262)	Data 0.001 (0.078)	Loss 0.975 (1.024)
Epoch: [36][200/200]	Time 0.170 (0.262)	Data 0.000 (0.077)	Loss 0.919 (1.031)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.071 (0.137)	Data 0.000 (0.055)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.069883108139038
==> Statistics for epoch 37: 573 clusters
Epoch: [37][20/200]	Time 0.187 (0.282)	Data 0.001 (0.100)	Loss 0.983 (0.307)
Epoch: [37][40/200]	Time 0.191 (0.270)	Data 0.001 (0.085)	Loss 1.314 (0.705)
Epoch: [37][60/200]	Time 0.183 (0.265)	Data 0.001 (0.080)	Loss 1.587 (0.856)
Epoch: [37][80/200]	Time 0.183 (0.261)	Data 0.000 (0.077)	Loss 1.113 (0.928)
Epoch: [37][100/200]	Time 0.174 (0.260)	Data 0.000 (0.076)	Loss 1.162 (0.964)
Epoch: [37][120/200]	Time 1.635 (0.268)	Data 1.291 (0.084)	Loss 1.137 (0.994)
Epoch: [37][140/200]	Time 0.174 (0.266)	Data 0.001 (0.083)	Loss 1.321 (1.011)
Epoch: [37][160/200]	Time 0.172 (0.265)	Data 0.001 (0.081)	Loss 1.171 (1.028)
Epoch: [37][180/200]	Time 0.171 (0.262)	Data 0.001 (0.079)	Loss 1.380 (1.043)
Epoch: [37][200/200]	Time 0.175 (0.260)	Data 0.000 (0.077)	Loss 1.291 (1.058)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.078 (0.134)	Data 0.000 (0.049)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.245232582092285
==> Statistics for epoch 38: 573 clusters
Epoch: [38][20/200]	Time 0.176 (0.286)	Data 0.001 (0.096)	Loss 1.098 (0.313)
Epoch: [38][40/200]	Time 0.166 (0.261)	Data 0.000 (0.078)	Loss 1.414 (0.724)
Epoch: [38][60/200]	Time 0.181 (0.255)	Data 0.001 (0.071)	Loss 1.198 (0.893)
Epoch: [38][80/200]	Time 0.182 (0.255)	Data 0.000 (0.071)	Loss 1.145 (0.939)
Epoch: [38][100/200]	Time 0.175 (0.254)	Data 0.000 (0.069)	Loss 1.253 (0.988)
Epoch: [38][120/200]	Time 1.535 (0.265)	Data 1.315 (0.081)	Loss 1.539 (1.022)
Epoch: [38][140/200]	Time 0.178 (0.263)	Data 0.001 (0.079)	Loss 1.193 (1.041)
Epoch: [38][160/200]	Time 0.179 (0.261)	Data 0.000 (0.077)	Loss 1.031 (1.056)
Epoch: [38][180/200]	Time 0.183 (0.260)	Data 0.000 (0.077)	Loss 1.108 (1.076)
Epoch: [38][200/200]	Time 0.178 (0.260)	Data 0.000 (0.076)	Loss 1.119 (1.086)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.082 (0.137)	Data 0.006 (0.054)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.165096759796143
==> Statistics for epoch 39: 573 clusters
Epoch: [39][20/200]	Time 0.180 (0.294)	Data 0.001 (0.105)	Loss 1.068 (0.283)
Epoch: [39][40/200]	Time 0.190 (0.270)	Data 0.001 (0.082)	Loss 1.235 (0.689)
Epoch: [39][60/200]	Time 0.187 (0.262)	Data 0.001 (0.079)	Loss 1.075 (0.840)
Epoch: [39][80/200]	Time 0.172 (0.260)	Data 0.000 (0.077)	Loss 1.314 (0.910)
Epoch: [39][100/200]	Time 0.286 (0.259)	Data 0.000 (0.075)	Loss 1.001 (0.945)
Epoch: [39][120/200]	Time 1.510 (0.268)	Data 1.273 (0.084)	Loss 1.040 (0.976)
Epoch: [39][140/200]	Time 0.173 (0.264)	Data 0.001 (0.081)	Loss 0.924 (1.009)
Epoch: [39][160/200]	Time 0.173 (0.263)	Data 0.001 (0.080)	Loss 0.889 (1.025)
Epoch: [39][180/200]	Time 0.182 (0.262)	Data 0.000 (0.079)	Loss 1.409 (1.049)
Epoch: [39][200/200]	Time 0.166 (0.262)	Data 0.000 (0.078)	Loss 1.166 (1.060)
Extract Features: [50/76]	Time 0.090 (0.136)	Data 0.007 (0.054)	
Mean AP: 87.9%

 * Finished epoch  39  model mAP: 87.9%  best: 87.9% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.140)	Data 0.000 (0.053)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.02271008491516
==> Statistics for epoch 40: 572 clusters
Epoch: [40][20/200]	Time 0.203 (0.291)	Data 0.000 (0.102)	Loss 1.130 (0.296)
Epoch: [40][40/200]	Time 0.178 (0.267)	Data 0.000 (0.082)	Loss 1.181 (0.702)
Epoch: [40][60/200]	Time 0.177 (0.264)	Data 0.001 (0.077)	Loss 1.346 (0.855)
Epoch: [40][80/200]	Time 0.177 (0.264)	Data 0.000 (0.075)	Loss 1.330 (0.934)
Epoch: [40][100/200]	Time 0.167 (0.260)	Data 0.000 (0.073)	Loss 1.145 (0.959)
Epoch: [40][120/200]	Time 1.434 (0.271)	Data 1.207 (0.084)	Loss 1.080 (0.998)
Epoch: [40][140/200]	Time 0.171 (0.269)	Data 0.001 (0.082)	Loss 1.264 (1.022)
Epoch: [40][160/200]	Time 0.183 (0.266)	Data 0.000 (0.080)	Loss 1.219 (1.043)
Epoch: [40][180/200]	Time 0.180 (0.265)	Data 0.001 (0.079)	Loss 1.039 (1.050)
Epoch: [40][200/200]	Time 0.177 (0.264)	Data 0.000 (0.077)	Loss 0.913 (1.059)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.136)	Data 0.000 (0.058)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.68201470375061
==> Statistics for epoch 41: 573 clusters
Epoch: [41][20/200]	Time 0.178 (0.273)	Data 0.001 (0.090)	Loss 1.278 (0.300)
Epoch: [41][40/200]	Time 0.181 (0.259)	Data 0.001 (0.075)	Loss 1.252 (0.704)
Epoch: [41][60/200]	Time 0.192 (0.256)	Data 0.001 (0.070)	Loss 1.231 (0.829)
Epoch: [41][80/200]	Time 0.175 (0.254)	Data 0.000 (0.068)	Loss 0.937 (0.891)
Epoch: [41][100/200]	Time 0.176 (0.251)	Data 0.000 (0.066)	Loss 0.961 (0.938)
Epoch: [41][120/200]	Time 1.464 (0.261)	Data 1.248 (0.076)	Loss 1.149 (0.959)
Epoch: [41][140/200]	Time 0.182 (0.257)	Data 0.001 (0.074)	Loss 1.513 (0.985)
Epoch: [41][160/200]	Time 0.177 (0.256)	Data 0.001 (0.072)	Loss 0.886 (1.002)
Epoch: [41][180/200]	Time 0.173 (0.253)	Data 0.001 (0.071)	Loss 1.264 (1.019)
Epoch: [41][200/200]	Time 0.168 (0.252)	Data 0.000 (0.070)	Loss 0.874 (1.030)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.077 (0.135)	Data 0.000 (0.051)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.9363751411438
==> Statistics for epoch 42: 574 clusters
Epoch: [42][20/200]	Time 0.187 (0.284)	Data 0.001 (0.093)	Loss 1.085 (0.331)
Epoch: [42][40/200]	Time 0.195 (0.262)	Data 0.000 (0.077)	Loss 1.573 (0.712)
Epoch: [42][60/200]	Time 0.190 (0.259)	Data 0.000 (0.073)	Loss 1.180 (0.858)
Epoch: [42][80/200]	Time 0.169 (0.256)	Data 0.000 (0.070)	Loss 1.509 (0.942)
Epoch: [42][100/200]	Time 0.174 (0.253)	Data 0.000 (0.068)	Loss 1.288 (1.000)
Epoch: [42][120/200]	Time 1.398 (0.263)	Data 1.180 (0.077)	Loss 1.486 (1.015)
Epoch: [42][140/200]	Time 0.175 (0.261)	Data 0.001 (0.076)	Loss 1.203 (1.026)
Epoch: [42][160/200]	Time 0.185 (0.259)	Data 0.001 (0.074)	Loss 1.174 (1.041)
Epoch: [42][180/200]	Time 0.172 (0.257)	Data 0.000 (0.073)	Loss 1.256 (1.060)
Epoch: [42][200/200]	Time 0.171 (0.256)	Data 0.000 (0.072)	Loss 0.566 (1.063)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.136)	Data 0.000 (0.057)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.964320182800293
==> Statistics for epoch 43: 574 clusters
Epoch: [43][20/200]	Time 0.188 (0.289)	Data 0.001 (0.100)	Loss 1.208 (0.297)
Epoch: [43][40/200]	Time 0.186 (0.269)	Data 0.001 (0.081)	Loss 1.069 (0.674)
Epoch: [43][60/200]	Time 0.181 (0.261)	Data 0.001 (0.074)	Loss 1.439 (0.850)
Epoch: [43][80/200]	Time 0.179 (0.256)	Data 0.000 (0.070)	Loss 0.827 (0.932)
Epoch: [43][100/200]	Time 0.169 (0.255)	Data 0.000 (0.069)	Loss 1.363 (0.993)
Epoch: [43][120/200]	Time 1.403 (0.264)	Data 1.161 (0.078)	Loss 1.145 (1.017)
Epoch: [43][140/200]	Time 0.192 (0.260)	Data 0.001 (0.075)	Loss 1.079 (1.027)
Epoch: [43][160/200]	Time 0.180 (0.259)	Data 0.001 (0.074)	Loss 1.293 (1.044)
Epoch: [43][180/200]	Time 0.181 (0.258)	Data 0.001 (0.072)	Loss 1.188 (1.050)
Epoch: [43][200/200]	Time 0.169 (0.256)	Data 0.000 (0.071)	Loss 0.954 (1.056)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.069 (0.136)	Data 0.000 (0.054)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.916536331176758
==> Statistics for epoch 44: 570 clusters
Epoch: [44][20/200]	Time 0.173 (0.290)	Data 0.001 (0.101)	Loss 1.069 (0.323)
Epoch: [44][40/200]	Time 0.186 (0.266)	Data 0.001 (0.080)	Loss 1.260 (0.723)
Epoch: [44][60/200]	Time 0.181 (0.263)	Data 0.001 (0.076)	Loss 1.185 (0.863)
Epoch: [44][80/200]	Time 0.181 (0.259)	Data 0.000 (0.073)	Loss 1.187 (0.926)
Epoch: [44][100/200]	Time 0.183 (0.256)	Data 0.000 (0.071)	Loss 1.090 (0.967)
Epoch: [44][120/200]	Time 1.451 (0.266)	Data 1.232 (0.081)	Loss 1.253 (0.976)
Epoch: [44][140/200]	Time 0.178 (0.262)	Data 0.000 (0.077)	Loss 0.921 (0.995)
Epoch: [44][160/200]	Time 0.178 (0.261)	Data 0.001 (0.075)	Loss 0.837 (1.004)
Epoch: [44][180/200]	Time 0.183 (0.259)	Data 0.001 (0.074)	Loss 1.011 (1.017)
Epoch: [44][200/200]	Time 0.294 (0.258)	Data 0.000 (0.072)	Loss 1.152 (1.026)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.135)	Data 0.000 (0.056)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.859033584594727
==> Statistics for epoch 45: 569 clusters
Epoch: [45][20/200]	Time 0.193 (0.283)	Data 0.001 (0.099)	Loss 1.059 (0.268)
Epoch: [45][40/200]	Time 0.197 (0.266)	Data 0.001 (0.081)	Loss 1.197 (0.645)
Epoch: [45][60/200]	Time 0.183 (0.258)	Data 0.001 (0.074)	Loss 1.070 (0.802)
Epoch: [45][80/200]	Time 0.176 (0.255)	Data 0.000 (0.070)	Loss 0.930 (0.885)
Epoch: [45][100/200]	Time 0.175 (0.253)	Data 0.000 (0.068)	Loss 1.111 (0.938)
Epoch: [45][120/200]	Time 1.353 (0.262)	Data 1.114 (0.076)	Loss 1.746 (0.963)
Epoch: [45][140/200]	Time 0.178 (0.259)	Data 0.001 (0.073)	Loss 1.015 (0.987)
Epoch: [45][160/200]	Time 0.187 (0.257)	Data 0.001 (0.072)	Loss 1.294 (1.010)
Epoch: [45][180/200]	Time 0.184 (0.256)	Data 0.001 (0.071)	Loss 1.344 (1.029)
Epoch: [45][200/200]	Time 0.167 (0.255)	Data 0.000 (0.070)	Loss 1.327 (1.043)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.137)	Data 0.000 (0.055)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.127701997756958
==> Statistics for epoch 46: 571 clusters
Epoch: [46][20/200]	Time 0.185 (0.282)	Data 0.001 (0.091)	Loss 1.180 (0.304)
Epoch: [46][40/200]	Time 0.176 (0.262)	Data 0.001 (0.076)	Loss 0.925 (0.701)
Epoch: [46][60/200]	Time 0.178 (0.257)	Data 0.001 (0.071)	Loss 1.504 (0.842)
Epoch: [46][80/200]	Time 0.174 (0.253)	Data 0.000 (0.069)	Loss 1.163 (0.914)
Epoch: [46][100/200]	Time 0.168 (0.251)	Data 0.000 (0.067)	Loss 0.939 (0.954)
Epoch: [46][120/200]	Time 1.381 (0.261)	Data 1.151 (0.076)	Loss 0.851 (0.980)
Epoch: [46][140/200]	Time 0.175 (0.261)	Data 0.001 (0.075)	Loss 1.004 (0.992)
Epoch: [46][160/200]	Time 0.181 (0.259)	Data 0.000 (0.074)	Loss 1.279 (1.023)
Epoch: [46][180/200]	Time 0.184 (0.257)	Data 0.000 (0.072)	Loss 0.850 (1.035)
Epoch: [46][200/200]	Time 0.175 (0.257)	Data 0.000 (0.072)	Loss 1.184 (1.049)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.070 (0.132)	Data 0.000 (0.049)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.223200798034668
==> Statistics for epoch 47: 571 clusters
Epoch: [47][20/200]	Time 0.186 (0.283)	Data 0.001 (0.091)	Loss 0.840 (0.294)
Epoch: [47][40/200]	Time 0.175 (0.264)	Data 0.001 (0.075)	Loss 0.897 (0.693)
Epoch: [47][60/200]	Time 0.184 (0.257)	Data 0.001 (0.071)	Loss 1.397 (0.856)
Epoch: [47][80/200]	Time 0.182 (0.254)	Data 0.000 (0.068)	Loss 0.955 (0.921)
Epoch: [47][100/200]	Time 0.290 (0.251)	Data 0.000 (0.065)	Loss 1.013 (0.957)
Epoch: [47][120/200]	Time 1.542 (0.261)	Data 1.313 (0.075)	Loss 0.946 (0.978)
Epoch: [47][140/200]	Time 0.181 (0.260)	Data 0.001 (0.074)	Loss 1.370 (0.994)
Epoch: [47][160/200]	Time 0.173 (0.259)	Data 0.001 (0.072)	Loss 1.272 (1.017)
Epoch: [47][180/200]	Time 0.178 (0.257)	Data 0.001 (0.071)	Loss 1.041 (1.029)
Epoch: [47][200/200]	Time 0.176 (0.256)	Data 0.000 (0.069)	Loss 0.996 (1.039)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.075 (0.138)	Data 0.000 (0.059)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.795899391174316
==> Statistics for epoch 48: 573 clusters
Epoch: [48][20/200]	Time 0.190 (0.276)	Data 0.001 (0.093)	Loss 1.138 (0.312)
Epoch: [48][40/200]	Time 0.174 (0.262)	Data 0.001 (0.077)	Loss 1.012 (0.684)
Epoch: [48][60/200]	Time 0.185 (0.255)	Data 0.001 (0.072)	Loss 1.075 (0.850)
Epoch: [48][80/200]	Time 0.169 (0.254)	Data 0.000 (0.069)	Loss 1.344 (0.910)
Epoch: [48][100/200]	Time 0.175 (0.253)	Data 0.000 (0.066)	Loss 1.198 (0.948)
Epoch: [48][120/200]	Time 1.382 (0.260)	Data 1.140 (0.075)	Loss 0.803 (0.970)
Epoch: [48][140/200]	Time 0.189 (0.258)	Data 0.001 (0.072)	Loss 1.085 (1.000)
Epoch: [48][160/200]	Time 0.179 (0.256)	Data 0.001 (0.071)	Loss 1.608 (1.011)
Epoch: [48][180/200]	Time 0.183 (0.255)	Data 0.001 (0.070)	Loss 1.206 (1.031)
Epoch: [48][200/200]	Time 0.173 (0.254)	Data 0.000 (0.069)	Loss 1.218 (1.046)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.076 (0.138)	Data 0.000 (0.056)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.99325203895569
==> Statistics for epoch 49: 572 clusters
Epoch: [49][20/200]	Time 0.192 (0.276)	Data 0.000 (0.091)	Loss 1.395 (0.320)
Epoch: [49][40/200]	Time 0.175 (0.260)	Data 0.001 (0.075)	Loss 1.131 (0.722)
Epoch: [49][60/200]	Time 0.202 (0.254)	Data 0.001 (0.070)	Loss 1.315 (0.852)
Epoch: [49][80/200]	Time 0.181 (0.253)	Data 0.000 (0.068)	Loss 1.070 (0.922)
Epoch: [49][100/200]	Time 0.175 (0.252)	Data 0.000 (0.067)	Loss 1.359 (0.960)
Epoch: [49][120/200]	Time 1.356 (0.262)	Data 1.106 (0.075)	Loss 1.291 (0.976)
Epoch: [49][140/200]	Time 0.180 (0.260)	Data 0.001 (0.073)	Loss 1.243 (1.001)
Epoch: [49][160/200]	Time 0.178 (0.259)	Data 0.001 (0.072)	Loss 0.862 (1.001)
Epoch: [49][180/200]	Time 0.184 (0.258)	Data 0.001 (0.071)	Loss 0.794 (1.011)
Epoch: [49][200/200]	Time 0.173 (0.256)	Data 0.000 (0.070)	Loss 1.167 (1.024)
Extract Features: [50/76]	Time 0.087 (0.133)	Data 0.000 (0.049)	
Mean AP: 88.0%

 * Finished epoch  49  model mAP: 88.0%  best: 88.0% *

==> Test with the best model:
=> Loaded checkpoint 'log/msmt2market/vit_tiny_cion/model_best.pth.tar'
Extract Features: [50/76]	Time 0.073 (0.134)	Data 0.000 (0.052)	
Mean AP: 88.0%
CMC Scores:
  top-1          94.7%
  top-5          97.9%
  top-10         98.5%
Total running time:  1:09:54.892910
