==========
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='resnet152', pretrained_path='/home/ma-user/work/Projects/ReIDNets_Checkpoints_TransReID/ResNet152.pth', features=0, dropout=0, momentum=0.2, drop_path_rate=0.3, hw_ratio=2, self_norm=True, conv_stem=False, 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/market/resnet152_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
  ----------------------------------------
pooling_type: gem
optimizer: SGD
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.561)	Data 0.000 (0.029)	
Computing jaccard distance...
Jaccard distance computing time cost: 21.35196852684021
Clustering criterion: eps: 0.600
==> Statistics for epoch 0: 613 clusters
Epoch: [0][20/200]	Time 1.770 (1.075)	Data 1.193 (0.092)	Loss 3.567 (2.723)
Epoch: [0][40/200]	Time 0.508 (0.830)	Data 0.001 (0.079)	Loss 2.732 (3.306)
Epoch: [0][60/200]	Time 0.511 (0.744)	Data 0.001 (0.073)	Loss 2.778 (3.175)
Epoch: [0][80/200]	Time 0.501 (0.703)	Data 0.001 (0.071)	Loss 3.118 (3.086)
Epoch: [0][100/200]	Time 0.505 (0.680)	Data 0.001 (0.069)	Loss 2.732 (2.988)
Epoch: [0][120/200]	Time 0.604 (0.665)	Data 0.001 (0.069)	Loss 2.775 (2.912)
Epoch: [0][140/200]	Time 0.500 (0.653)	Data 0.001 (0.068)	Loss 2.763 (2.851)
Epoch: [0][160/200]	Time 0.498 (0.644)	Data 0.001 (0.067)	Loss 1.713 (2.784)
Epoch: [0][180/200]	Time 0.499 (0.638)	Data 0.001 (0.067)	Loss 2.140 (2.722)
Epoch: [0][200/200]	Time 0.622 (0.633)	Data 0.001 (0.067)	Loss 2.337 (2.677)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.183 (0.225)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.108896255493164
==> Statistics for epoch 1: 584 clusters
Epoch: [1][20/200]	Time 0.620 (0.628)	Data 0.001 (0.100)	Loss 2.682 (1.104)
Epoch: [1][40/200]	Time 0.503 (0.607)	Data 0.001 (0.085)	Loss 1.987 (1.679)
Epoch: [1][60/200]	Time 0.505 (0.600)	Data 0.001 (0.080)	Loss 2.422 (1.799)
Epoch: [1][80/200]	Time 0.589 (0.599)	Data 0.001 (0.079)	Loss 2.134 (1.844)
Epoch: [1][100/200]	Time 0.501 (0.598)	Data 0.001 (0.077)	Loss 2.200 (1.865)
Epoch: [1][120/200]	Time 0.502 (0.597)	Data 0.000 (0.075)	Loss 1.968 (1.885)
Epoch: [1][140/200]	Time 0.502 (0.595)	Data 0.000 (0.074)	Loss 2.297 (1.888)
Epoch: [1][160/200]	Time 0.575 (0.595)	Data 0.000 (0.073)	Loss 1.855 (1.880)
Epoch: [1][180/200]	Time 0.501 (0.593)	Data 0.000 (0.072)	Loss 1.869 (1.874)
Epoch: [1][200/200]	Time 0.526 (0.599)	Data 0.001 (0.078)	Loss 1.665 (1.867)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.183 (0.227)	Data 0.000 (0.028)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.599430084228516
==> Statistics for epoch 2: 593 clusters
Epoch: [2][20/200]	Time 0.504 (0.643)	Data 0.001 (0.109)	Loss 2.082 (0.527)
Epoch: [2][40/200]	Time 0.515 (0.623)	Data 0.002 (0.091)	Loss 1.968 (1.158)
Epoch: [2][60/200]	Time 0.504 (0.613)	Data 0.001 (0.085)	Loss 1.509 (1.351)
Epoch: [2][80/200]	Time 0.607 (0.613)	Data 0.001 (0.086)	Loss 1.817 (1.466)
Epoch: [2][100/200]	Time 0.502 (0.609)	Data 0.001 (0.083)	Loss 1.641 (1.522)
Epoch: [2][120/200]	Time 0.506 (0.608)	Data 0.000 (0.081)	Loss 1.569 (1.544)
Epoch: [2][140/200]	Time 0.503 (0.606)	Data 0.000 (0.081)	Loss 1.516 (1.561)
Epoch: [2][160/200]	Time 0.504 (0.605)	Data 0.000 (0.080)	Loss 1.850 (1.580)
Epoch: [2][180/200]	Time 0.505 (0.605)	Data 0.000 (0.080)	Loss 1.676 (1.579)
Epoch: [2][200/200]	Time 0.508 (0.610)	Data 0.001 (0.085)	Loss 1.951 (1.577)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.273 (0.226)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.61240816116333
==> Statistics for epoch 3: 589 clusters
Epoch: [3][20/200]	Time 0.502 (0.636)	Data 0.001 (0.112)	Loss 1.813 (0.480)
Epoch: [3][40/200]	Time 0.506 (0.621)	Data 0.001 (0.094)	Loss 1.403 (1.066)
Epoch: [3][60/200]	Time 0.501 (0.614)	Data 0.001 (0.089)	Loss 1.829 (1.258)
Epoch: [3][80/200]	Time 0.507 (0.608)	Data 0.001 (0.085)	Loss 1.730 (1.344)
Epoch: [3][100/200]	Time 0.600 (0.606)	Data 0.001 (0.082)	Loss 1.927 (1.400)
Epoch: [3][120/200]	Time 0.506 (0.605)	Data 0.000 (0.082)	Loss 1.584 (1.426)
Epoch: [3][140/200]	Time 0.501 (0.605)	Data 0.000 (0.081)	Loss 1.583 (1.440)
Epoch: [3][160/200]	Time 0.511 (0.604)	Data 0.000 (0.079)	Loss 1.734 (1.457)
Epoch: [3][180/200]	Time 0.508 (0.603)	Data 0.000 (0.078)	Loss 1.530 (1.464)
Epoch: [3][200/200]	Time 0.508 (0.610)	Data 0.001 (0.086)	Loss 1.553 (1.464)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.181 (0.223)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.64075803756714
==> Statistics for epoch 4: 586 clusters
Epoch: [4][20/200]	Time 0.504 (0.633)	Data 0.001 (0.117)	Loss 1.174 (0.398)
Epoch: [4][40/200]	Time 0.510 (0.626)	Data 0.001 (0.104)	Loss 1.285 (0.950)
Epoch: [4][60/200]	Time 0.620 (0.619)	Data 0.001 (0.095)	Loss 1.136 (1.103)
Epoch: [4][80/200]	Time 0.509 (0.615)	Data 0.002 (0.091)	Loss 1.276 (1.187)
Epoch: [4][100/200]	Time 0.507 (0.613)	Data 0.001 (0.088)	Loss 1.646 (1.226)
Epoch: [4][120/200]	Time 0.515 (0.610)	Data 0.000 (0.085)	Loss 1.374 (1.255)
Epoch: [4][140/200]	Time 0.504 (0.609)	Data 0.000 (0.084)	Loss 1.390 (1.264)
Epoch: [4][160/200]	Time 0.505 (0.607)	Data 0.000 (0.081)	Loss 1.352 (1.273)
Epoch: [4][180/200]	Time 0.503 (0.606)	Data 0.000 (0.080)	Loss 1.350 (1.281)
Epoch: [4][200/200]	Time 0.509 (0.612)	Data 0.001 (0.086)	Loss 1.285 (1.281)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.182 (0.214)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.051658868789673
==> Statistics for epoch 5: 599 clusters
Epoch: [5][20/200]	Time 0.503 (0.627)	Data 0.001 (0.107)	Loss 1.809 (0.376)
Epoch: [5][40/200]	Time 0.504 (0.611)	Data 0.001 (0.088)	Loss 1.053 (0.817)
Epoch: [5][60/200]	Time 0.503 (0.607)	Data 0.001 (0.083)	Loss 1.504 (0.994)
Epoch: [5][80/200]	Time 0.507 (0.601)	Data 0.001 (0.078)	Loss 1.395 (1.100)
Epoch: [5][100/200]	Time 0.505 (0.599)	Data 0.001 (0.076)	Loss 1.814 (1.131)
Epoch: [5][120/200]	Time 0.505 (0.597)	Data 0.000 (0.074)	Loss 1.076 (1.168)
Epoch: [5][140/200]	Time 0.516 (0.597)	Data 0.000 (0.073)	Loss 1.594 (1.188)
Epoch: [5][160/200]	Time 0.509 (0.595)	Data 0.000 (0.072)	Loss 1.255 (1.204)
Epoch: [5][180/200]	Time 0.611 (0.596)	Data 0.000 (0.072)	Loss 1.381 (1.213)
Epoch: [5][200/200]	Time 0.507 (0.602)	Data 0.001 (0.078)	Loss 1.263 (1.221)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.182 (0.218)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.171994924545288
==> Statistics for epoch 6: 606 clusters
Epoch: [6][20/200]	Time 0.503 (0.620)	Data 0.001 (0.096)	Loss 0.984 (0.328)
Epoch: [6][40/200]	Time 0.505 (0.603)	Data 0.001 (0.081)	Loss 1.147 (0.753)
Epoch: [6][60/200]	Time 0.521 (0.601)	Data 0.001 (0.078)	Loss 1.017 (0.903)
Epoch: [6][80/200]	Time 0.633 (0.599)	Data 0.001 (0.075)	Loss 1.046 (0.980)
Epoch: [6][100/200]	Time 0.514 (0.597)	Data 0.002 (0.073)	Loss 1.073 (1.034)
Epoch: [6][120/200]	Time 0.600 (0.597)	Data 0.000 (0.072)	Loss 1.061 (1.065)
Epoch: [6][140/200]	Time 0.500 (0.596)	Data 0.000 (0.071)	Loss 0.989 (1.085)
Epoch: [6][160/200]	Time 0.506 (0.596)	Data 0.000 (0.069)	Loss 1.068 (1.108)
Epoch: [6][180/200]	Time 0.505 (0.595)	Data 0.000 (0.070)	Loss 1.313 (1.119)
Epoch: [6][200/200]	Time 0.506 (0.603)	Data 0.002 (0.077)	Loss 1.134 (1.127)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.181 (0.221)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.371604204177856
==> Statistics for epoch 7: 602 clusters
Epoch: [7][20/200]	Time 0.513 (0.633)	Data 0.001 (0.112)	Loss 1.249 (0.339)
Epoch: [7][40/200]	Time 0.506 (0.624)	Data 0.001 (0.099)	Loss 1.129 (0.723)
Epoch: [7][60/200]	Time 0.504 (0.613)	Data 0.004 (0.089)	Loss 1.355 (0.873)
Epoch: [7][80/200]	Time 0.519 (0.606)	Data 0.001 (0.084)	Loss 1.104 (0.931)
Epoch: [7][100/200]	Time 0.506 (0.601)	Data 0.001 (0.079)	Loss 1.390 (0.964)
Epoch: [7][120/200]	Time 0.502 (0.600)	Data 0.000 (0.078)	Loss 0.832 (0.992)
Epoch: [7][140/200]	Time 0.577 (0.598)	Data 0.000 (0.075)	Loss 1.076 (1.017)
Epoch: [7][160/200]	Time 0.504 (0.596)	Data 0.000 (0.074)	Loss 0.980 (1.033)
Epoch: [7][180/200]	Time 0.592 (0.595)	Data 0.000 (0.073)	Loss 1.000 (1.045)
Epoch: [7][200/200]	Time 0.499 (0.601)	Data 0.001 (0.080)	Loss 1.224 (1.052)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.256 (0.218)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.27349305152893
==> Statistics for epoch 8: 611 clusters
Epoch: [8][20/200]	Time 1.941 (0.614)	Data 1.243 (0.095)	Loss 0.809 (0.201)
Epoch: [8][40/200]	Time 0.512 (0.596)	Data 0.001 (0.078)	Loss 1.256 (0.634)
Epoch: [8][60/200]	Time 0.527 (0.594)	Data 0.001 (0.072)	Loss 1.193 (0.813)
Epoch: [8][80/200]	Time 0.651 (0.598)	Data 0.002 (0.073)	Loss 0.956 (0.886)
Epoch: [8][100/200]	Time 0.502 (0.598)	Data 0.001 (0.073)	Loss 1.172 (0.940)
Epoch: [8][120/200]	Time 0.608 (0.597)	Data 0.001 (0.072)	Loss 1.436 (0.982)
Epoch: [8][140/200]	Time 0.503 (0.595)	Data 0.001 (0.071)	Loss 0.845 (0.979)
Epoch: [8][160/200]	Time 0.507 (0.594)	Data 0.001 (0.070)	Loss 0.681 (0.993)
Epoch: [8][180/200]	Time 0.522 (0.592)	Data 0.001 (0.069)	Loss 1.280 (1.008)
Epoch: [8][200/200]	Time 0.580 (0.592)	Data 0.001 (0.069)	Loss 0.979 (1.021)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.184 (0.218)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.031044960021973
==> Statistics for epoch 9: 609 clusters
Epoch: [9][20/200]	Time 1.793 (0.626)	Data 1.252 (0.099)	Loss 0.969 (0.257)
Epoch: [9][40/200]	Time 0.524 (0.604)	Data 0.001 (0.081)	Loss 0.744 (0.640)
Epoch: [9][60/200]	Time 0.604 (0.599)	Data 0.001 (0.076)	Loss 1.162 (0.773)
Epoch: [9][80/200]	Time 0.519 (0.596)	Data 0.001 (0.074)	Loss 1.324 (0.846)
Epoch: [9][100/200]	Time 0.509 (0.594)	Data 0.001 (0.073)	Loss 1.204 (0.891)
Epoch: [9][120/200]	Time 0.508 (0.592)	Data 0.001 (0.070)	Loss 1.293 (0.921)
Epoch: [9][140/200]	Time 0.512 (0.591)	Data 0.001 (0.069)	Loss 0.953 (0.941)
Epoch: [9][160/200]	Time 0.586 (0.591)	Data 0.001 (0.068)	Loss 1.095 (0.950)
Epoch: [9][180/200]	Time 0.506 (0.590)	Data 0.001 (0.068)	Loss 1.102 (0.959)
Epoch: [9][200/200]	Time 0.587 (0.589)	Data 0.001 (0.067)	Loss 1.071 (0.969)
Extract Features: [50/76]	Time 0.348 (0.218)	Data 0.000 (0.020)	
Mean AP: 91.7%

 * Finished epoch   9  model mAP: 91.7%  best: 91.7% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.218)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.906195163726807
==> Statistics for epoch 10: 604 clusters
Epoch: [10][20/200]	Time 0.504 (0.690)	Data 0.001 (0.106)	Loss 1.180 (0.268)
Epoch: [10][40/200]	Time 0.504 (0.635)	Data 0.001 (0.084)	Loss 1.075 (0.627)
Epoch: [10][60/200]	Time 0.611 (0.623)	Data 0.001 (0.080)	Loss 1.188 (0.777)
Epoch: [10][80/200]	Time 0.617 (0.613)	Data 0.001 (0.075)	Loss 0.849 (0.842)
Epoch: [10][100/200]	Time 0.518 (0.608)	Data 0.002 (0.074)	Loss 0.893 (0.891)
Epoch: [10][120/200]	Time 0.510 (0.606)	Data 0.000 (0.074)	Loss 0.987 (0.904)
Epoch: [10][140/200]	Time 0.577 (0.604)	Data 0.000 (0.073)	Loss 1.010 (0.914)
Epoch: [10][160/200]	Time 0.506 (0.602)	Data 0.000 (0.072)	Loss 1.019 (0.925)
Epoch: [10][180/200]	Time 0.571 (0.600)	Data 0.000 (0.071)	Loss 1.035 (0.932)
Epoch: [10][200/200]	Time 0.507 (0.605)	Data 0.001 (0.077)	Loss 1.023 (0.940)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.257 (0.216)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.000503540039062
==> Statistics for epoch 11: 604 clusters
Epoch: [11][20/200]	Time 0.512 (0.624)	Data 0.001 (0.099)	Loss 1.100 (0.269)
Epoch: [11][40/200]	Time 0.605 (0.600)	Data 0.001 (0.081)	Loss 0.908 (0.606)
Epoch: [11][60/200]	Time 0.515 (0.600)	Data 0.001 (0.079)	Loss 0.874 (0.722)
Epoch: [11][80/200]	Time 0.511 (0.599)	Data 0.001 (0.075)	Loss 1.341 (0.775)
Epoch: [11][100/200]	Time 0.587 (0.597)	Data 0.001 (0.073)	Loss 0.678 (0.812)
Epoch: [11][120/200]	Time 0.501 (0.595)	Data 0.000 (0.072)	Loss 1.003 (0.840)
Epoch: [11][140/200]	Time 0.501 (0.594)	Data 0.000 (0.071)	Loss 0.970 (0.872)
Epoch: [11][160/200]	Time 0.503 (0.593)	Data 0.000 (0.070)	Loss 0.820 (0.883)
Epoch: [11][180/200]	Time 0.504 (0.591)	Data 0.000 (0.069)	Loss 1.074 (0.892)
Epoch: [11][200/200]	Time 0.507 (0.597)	Data 0.001 (0.075)	Loss 1.078 (0.894)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.215)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.30873417854309
==> Statistics for epoch 12: 610 clusters
Epoch: [12][20/200]	Time 1.920 (0.626)	Data 1.387 (0.105)	Loss 0.733 (0.192)
Epoch: [12][40/200]	Time 0.507 (0.607)	Data 0.001 (0.088)	Loss 1.070 (0.586)
Epoch: [12][60/200]	Time 0.512 (0.605)	Data 0.001 (0.083)	Loss 1.283 (0.707)
Epoch: [12][80/200]	Time 0.515 (0.603)	Data 0.002 (0.078)	Loss 0.768 (0.768)
Epoch: [12][100/200]	Time 0.508 (0.599)	Data 0.001 (0.075)	Loss 0.876 (0.801)
Epoch: [12][120/200]	Time 0.622 (0.597)	Data 0.002 (0.073)	Loss 0.737 (0.819)
Epoch: [12][140/200]	Time 0.508 (0.595)	Data 0.001 (0.071)	Loss 1.127 (0.838)
Epoch: [12][160/200]	Time 0.508 (0.596)	Data 0.001 (0.072)	Loss 0.965 (0.848)
Epoch: [12][180/200]	Time 0.597 (0.596)	Data 0.001 (0.071)	Loss 1.034 (0.861)
Epoch: [12][200/200]	Time 0.505 (0.595)	Data 0.001 (0.070)	Loss 0.975 (0.869)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.216)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.49034023284912
==> Statistics for epoch 13: 603 clusters
Epoch: [13][20/200]	Time 0.509 (0.637)	Data 0.001 (0.109)	Loss 1.244 (0.256)
Epoch: [13][40/200]	Time 0.635 (0.614)	Data 0.001 (0.086)	Loss 1.045 (0.554)
Epoch: [13][60/200]	Time 0.501 (0.604)	Data 0.001 (0.079)	Loss 1.312 (0.666)
Epoch: [13][80/200]	Time 0.510 (0.599)	Data 0.001 (0.075)	Loss 0.990 (0.726)
Epoch: [13][100/200]	Time 0.510 (0.598)	Data 0.002 (0.073)	Loss 1.490 (0.775)
Epoch: [13][120/200]	Time 0.509 (0.599)	Data 0.000 (0.074)	Loss 0.961 (0.789)
Epoch: [13][140/200]	Time 0.506 (0.597)	Data 0.000 (0.072)	Loss 0.455 (0.801)
Epoch: [13][160/200]	Time 0.504 (0.596)	Data 0.000 (0.071)	Loss 1.016 (0.816)
Epoch: [13][180/200]	Time 0.586 (0.596)	Data 0.000 (0.070)	Loss 0.934 (0.820)
Epoch: [13][200/200]	Time 0.504 (0.602)	Data 0.001 (0.077)	Loss 0.738 (0.823)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.181 (0.214)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.55013918876648
==> Statistics for epoch 14: 603 clusters
Epoch: [14][20/200]	Time 0.502 (0.628)	Data 0.001 (0.103)	Loss 0.831 (0.223)
Epoch: [14][40/200]	Time 0.518 (0.612)	Data 0.001 (0.087)	Loss 1.038 (0.528)
Epoch: [14][60/200]	Time 0.506 (0.608)	Data 0.003 (0.083)	Loss 0.636 (0.637)
Epoch: [14][80/200]	Time 0.500 (0.603)	Data 0.001 (0.079)	Loss 0.681 (0.690)
Epoch: [14][100/200]	Time 0.505 (0.600)	Data 0.001 (0.075)	Loss 0.845 (0.720)
Epoch: [14][120/200]	Time 0.584 (0.597)	Data 0.000 (0.073)	Loss 0.898 (0.745)
Epoch: [14][140/200]	Time 0.593 (0.596)	Data 0.000 (0.072)	Loss 0.919 (0.755)
Epoch: [14][160/200]	Time 0.505 (0.595)	Data 0.000 (0.071)	Loss 0.682 (0.759)
Epoch: [14][180/200]	Time 0.499 (0.594)	Data 0.000 (0.070)	Loss 0.943 (0.776)
Epoch: [14][200/200]	Time 0.629 (0.600)	Data 0.001 (0.076)	Loss 1.220 (0.789)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.215)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.314739227294922
==> Statistics for epoch 15: 605 clusters
Epoch: [15][20/200]	Time 0.546 (0.649)	Data 0.002 (0.112)	Loss 0.897 (0.234)
Epoch: [15][40/200]	Time 0.618 (0.622)	Data 0.002 (0.090)	Loss 0.405 (0.522)
Epoch: [15][60/200]	Time 0.500 (0.614)	Data 0.001 (0.086)	Loss 0.952 (0.638)
Epoch: [15][80/200]	Time 0.505 (0.606)	Data 0.001 (0.081)	Loss 0.738 (0.699)
Epoch: [15][100/200]	Time 0.503 (0.601)	Data 0.001 (0.077)	Loss 0.648 (0.718)
Epoch: [15][120/200]	Time 0.613 (0.601)	Data 0.000 (0.077)	Loss 1.116 (0.737)
Epoch: [15][140/200]	Time 0.503 (0.599)	Data 0.000 (0.075)	Loss 0.911 (0.747)
Epoch: [15][160/200]	Time 0.507 (0.599)	Data 0.000 (0.074)	Loss 1.063 (0.759)
Epoch: [15][180/200]	Time 0.505 (0.598)	Data 0.000 (0.074)	Loss 0.781 (0.771)
Epoch: [15][200/200]	Time 0.525 (0.605)	Data 0.001 (0.080)	Loss 0.705 (0.776)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.215)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.029293298721313
==> Statistics for epoch 16: 600 clusters
Epoch: [16][20/200]	Time 0.504 (0.634)	Data 0.001 (0.108)	Loss 1.043 (0.242)
Epoch: [16][40/200]	Time 0.502 (0.613)	Data 0.001 (0.092)	Loss 0.836 (0.539)
Epoch: [16][60/200]	Time 0.511 (0.604)	Data 0.002 (0.082)	Loss 0.601 (0.628)
Epoch: [16][80/200]	Time 0.505 (0.606)	Data 0.001 (0.081)	Loss 0.702 (0.672)
Epoch: [16][100/200]	Time 0.503 (0.604)	Data 0.001 (0.079)	Loss 0.642 (0.692)
Epoch: [16][120/200]	Time 0.511 (0.601)	Data 0.000 (0.076)	Loss 0.666 (0.713)
Epoch: [16][140/200]	Time 0.502 (0.602)	Data 0.000 (0.076)	Loss 1.244 (0.735)
Epoch: [16][160/200]	Time 0.501 (0.600)	Data 0.000 (0.075)	Loss 0.645 (0.743)
Epoch: [16][180/200]	Time 0.501 (0.598)	Data 0.000 (0.074)	Loss 0.874 (0.753)
Epoch: [16][200/200]	Time 0.510 (0.603)	Data 0.001 (0.079)	Loss 0.851 (0.762)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.256 (0.216)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.63615322113037
==> Statistics for epoch 17: 603 clusters
Epoch: [17][20/200]	Time 0.501 (0.629)	Data 0.001 (0.103)	Loss 0.691 (0.189)
Epoch: [17][40/200]	Time 0.501 (0.608)	Data 0.001 (0.083)	Loss 0.613 (0.475)
Epoch: [17][60/200]	Time 0.503 (0.607)	Data 0.001 (0.081)	Loss 0.543 (0.564)
Epoch: [17][80/200]	Time 0.505 (0.604)	Data 0.001 (0.080)	Loss 0.956 (0.625)
Epoch: [17][100/200]	Time 0.507 (0.603)	Data 0.001 (0.078)	Loss 0.658 (0.669)
Epoch: [17][120/200]	Time 0.506 (0.603)	Data 0.000 (0.078)	Loss 0.821 (0.682)
Epoch: [17][140/200]	Time 0.501 (0.602)	Data 0.000 (0.078)	Loss 0.701 (0.691)
Epoch: [17][160/200]	Time 0.508 (0.601)	Data 0.000 (0.076)	Loss 0.819 (0.707)
Epoch: [17][180/200]	Time 0.501 (0.601)	Data 0.000 (0.076)	Loss 0.766 (0.713)
Epoch: [17][200/200]	Time 0.520 (0.608)	Data 0.001 (0.083)	Loss 0.923 (0.719)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.178 (0.218)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.317795991897583
==> Statistics for epoch 18: 599 clusters
Epoch: [18][20/200]	Time 0.612 (0.632)	Data 0.001 (0.109)	Loss 0.706 (0.189)
Epoch: [18][40/200]	Time 0.521 (0.608)	Data 0.001 (0.086)	Loss 1.106 (0.467)
Epoch: [18][60/200]	Time 0.501 (0.603)	Data 0.001 (0.081)	Loss 1.160 (0.581)
Epoch: [18][80/200]	Time 0.500 (0.601)	Data 0.001 (0.079)	Loss 1.095 (0.648)
Epoch: [18][100/200]	Time 0.504 (0.600)	Data 0.001 (0.079)	Loss 0.835 (0.670)
Epoch: [18][120/200]	Time 0.501 (0.598)	Data 0.000 (0.076)	Loss 0.836 (0.688)
Epoch: [18][140/200]	Time 0.504 (0.596)	Data 0.000 (0.075)	Loss 0.829 (0.705)
Epoch: [18][160/200]	Time 0.501 (0.596)	Data 0.000 (0.074)	Loss 0.806 (0.717)
Epoch: [18][180/200]	Time 0.607 (0.596)	Data 0.000 (0.073)	Loss 0.654 (0.727)
Epoch: [18][200/200]	Time 0.541 (0.603)	Data 0.001 (0.080)	Loss 0.439 (0.728)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.216)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.57972812652588
==> Statistics for epoch 19: 602 clusters
Epoch: [19][20/200]	Time 0.509 (0.630)	Data 0.001 (0.101)	Loss 0.692 (0.155)
Epoch: [19][40/200]	Time 0.613 (0.606)	Data 0.001 (0.082)	Loss 0.930 (0.432)
Epoch: [19][60/200]	Time 0.513 (0.594)	Data 0.001 (0.075)	Loss 0.708 (0.532)
Epoch: [19][80/200]	Time 0.505 (0.591)	Data 0.001 (0.073)	Loss 0.457 (0.596)
Epoch: [19][100/200]	Time 0.500 (0.591)	Data 0.001 (0.071)	Loss 0.781 (0.627)
Epoch: [19][120/200]	Time 0.508 (0.590)	Data 0.000 (0.070)	Loss 0.792 (0.651)
Epoch: [19][140/200]	Time 0.620 (0.592)	Data 0.000 (0.071)	Loss 0.889 (0.671)
Epoch: [19][160/200]	Time 0.500 (0.591)	Data 0.000 (0.070)	Loss 0.953 (0.686)
Epoch: [19][180/200]	Time 0.499 (0.590)	Data 0.000 (0.069)	Loss 0.671 (0.693)
Epoch: [19][200/200]	Time 0.604 (0.597)	Data 0.001 (0.075)	Loss 0.784 (0.695)
Extract Features: [50/76]	Time 0.299 (0.219)	Data 0.000 (0.022)	
Mean AP: 92.7%

 * Finished epoch  19  model mAP: 92.7%  best: 92.7% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.214)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.111589670181274
==> Statistics for epoch 20: 600 clusters
Epoch: [20][20/200]	Time 0.505 (0.635)	Data 0.000 (0.115)	Loss 0.636 (0.150)
Epoch: [20][40/200]	Time 0.513 (0.609)	Data 0.000 (0.092)	Loss 0.770 (0.404)
Epoch: [20][60/200]	Time 0.515 (0.603)	Data 0.000 (0.083)	Loss 0.525 (0.489)
Epoch: [20][80/200]	Time 0.587 (0.600)	Data 0.000 (0.079)	Loss 0.558 (0.536)
Epoch: [20][100/200]	Time 0.509 (0.596)	Data 0.000 (0.076)	Loss 0.485 (0.574)
Epoch: [20][120/200]	Time 0.498 (0.594)	Data 0.000 (0.074)	Loss 0.733 (0.597)
Epoch: [20][140/200]	Time 0.501 (0.594)	Data 0.000 (0.074)	Loss 0.664 (0.608)
Epoch: [20][160/200]	Time 0.502 (0.593)	Data 0.000 (0.073)	Loss 0.863 (0.628)
Epoch: [20][180/200]	Time 0.501 (0.593)	Data 0.000 (0.073)	Loss 0.603 (0.639)
Epoch: [20][200/200]	Time 0.501 (0.600)	Data 0.000 (0.079)	Loss 0.597 (0.647)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.215)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.060808897018433
==> Statistics for epoch 21: 600 clusters
Epoch: [21][20/200]	Time 0.500 (0.634)	Data 0.001 (0.106)	Loss 0.746 (0.160)
Epoch: [21][40/200]	Time 0.505 (0.611)	Data 0.002 (0.087)	Loss 0.684 (0.427)
Epoch: [21][60/200]	Time 0.512 (0.610)	Data 0.001 (0.083)	Loss 0.729 (0.525)
Epoch: [21][80/200]	Time 0.505 (0.604)	Data 0.001 (0.079)	Loss 0.630 (0.577)
Epoch: [21][100/200]	Time 0.509 (0.600)	Data 0.001 (0.076)	Loss 0.608 (0.612)
Epoch: [21][120/200]	Time 0.507 (0.599)	Data 0.000 (0.075)	Loss 0.912 (0.624)
Epoch: [21][140/200]	Time 0.503 (0.600)	Data 0.000 (0.074)	Loss 0.758 (0.627)
Epoch: [21][160/200]	Time 0.506 (0.598)	Data 0.000 (0.074)	Loss 0.746 (0.638)
Epoch: [21][180/200]	Time 0.510 (0.598)	Data 0.000 (0.073)	Loss 0.628 (0.647)
Epoch: [21][200/200]	Time 0.510 (0.606)	Data 0.001 (0.080)	Loss 0.588 (0.654)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.216)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.16498851776123
==> Statistics for epoch 22: 602 clusters
Epoch: [22][20/200]	Time 0.628 (0.638)	Data 0.001 (0.112)	Loss 0.602 (0.178)
Epoch: [22][40/200]	Time 0.525 (0.611)	Data 0.002 (0.088)	Loss 0.633 (0.435)
Epoch: [22][60/200]	Time 0.506 (0.604)	Data 0.001 (0.081)	Loss 0.630 (0.540)
Epoch: [22][80/200]	Time 0.623 (0.600)	Data 0.002 (0.077)	Loss 0.539 (0.573)
Epoch: [22][100/200]	Time 0.511 (0.599)	Data 0.002 (0.075)	Loss 0.790 (0.599)
Epoch: [22][120/200]	Time 0.504 (0.598)	Data 0.000 (0.073)	Loss 0.937 (0.620)
Epoch: [22][140/200]	Time 0.509 (0.597)	Data 0.001 (0.072)	Loss 0.486 (0.625)
Epoch: [22][160/200]	Time 0.502 (0.596)	Data 0.000 (0.072)	Loss 0.835 (0.633)
Epoch: [22][180/200]	Time 0.503 (0.594)	Data 0.000 (0.071)	Loss 0.667 (0.633)
Epoch: [22][200/200]	Time 0.506 (0.600)	Data 0.001 (0.077)	Loss 0.574 (0.634)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.214)	Data 0.000 (0.020)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.582077503204346
==> Statistics for epoch 23: 599 clusters
Epoch: [23][20/200]	Time 0.515 (0.624)	Data 0.001 (0.095)	Loss 0.496 (0.147)
Epoch: [23][40/200]	Time 0.610 (0.609)	Data 0.001 (0.081)	Loss 0.646 (0.408)
Epoch: [23][60/200]	Time 0.512 (0.599)	Data 0.001 (0.074)	Loss 0.630 (0.519)
Epoch: [23][80/200]	Time 0.633 (0.602)	Data 0.000 (0.074)	Loss 0.693 (0.572)
Epoch: [23][100/200]	Time 0.502 (0.598)	Data 0.001 (0.072)	Loss 1.150 (0.592)
Epoch: [23][120/200]	Time 0.501 (0.596)	Data 0.000 (0.071)	Loss 0.700 (0.602)
Epoch: [23][140/200]	Time 0.501 (0.595)	Data 0.000 (0.070)	Loss 0.672 (0.620)
Epoch: [23][160/200]	Time 0.505 (0.594)	Data 0.000 (0.070)	Loss 0.594 (0.629)
Epoch: [23][180/200]	Time 0.498 (0.593)	Data 0.000 (0.069)	Loss 0.830 (0.631)
Epoch: [23][200/200]	Time 0.504 (0.601)	Data 0.001 (0.077)	Loss 0.655 (0.630)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.178 (0.215)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.103037118911743
==> Statistics for epoch 24: 602 clusters
Epoch: [24][20/200]	Time 0.498 (0.634)	Data 0.001 (0.105)	Loss 0.506 (0.154)
Epoch: [24][40/200]	Time 0.610 (0.611)	Data 0.001 (0.086)	Loss 0.913 (0.422)
Epoch: [24][60/200]	Time 0.508 (0.601)	Data 0.001 (0.079)	Loss 0.668 (0.499)
Epoch: [24][80/200]	Time 0.501 (0.597)	Data 0.000 (0.075)	Loss 0.665 (0.572)
Epoch: [24][100/200]	Time 0.516 (0.595)	Data 0.001 (0.074)	Loss 0.728 (0.582)
Epoch: [24][120/200]	Time 0.501 (0.594)	Data 0.000 (0.072)	Loss 0.737 (0.598)
Epoch: [24][140/200]	Time 0.508 (0.594)	Data 0.000 (0.071)	Loss 0.715 (0.599)
Epoch: [24][160/200]	Time 0.591 (0.593)	Data 0.000 (0.071)	Loss 0.455 (0.609)
Epoch: [24][180/200]	Time 0.509 (0.592)	Data 0.000 (0.070)	Loss 0.776 (0.617)
Epoch: [24][200/200]	Time 0.524 (0.598)	Data 0.000 (0.075)	Loss 0.760 (0.622)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.178 (0.212)	Data 0.000 (0.020)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.435568809509277
==> Statistics for epoch 25: 602 clusters
Epoch: [25][20/200]	Time 0.504 (0.637)	Data 0.001 (0.099)	Loss 0.539 (0.166)
Epoch: [25][40/200]	Time 0.604 (0.611)	Data 0.001 (0.083)	Loss 0.555 (0.419)
Epoch: [25][60/200]	Time 0.503 (0.602)	Data 0.001 (0.077)	Loss 0.926 (0.514)
Epoch: [25][80/200]	Time 0.527 (0.598)	Data 0.002 (0.074)	Loss 0.751 (0.569)
Epoch: [25][100/200]	Time 0.634 (0.597)	Data 0.001 (0.073)	Loss 0.828 (0.593)
Epoch: [25][120/200]	Time 0.502 (0.596)	Data 0.000 (0.073)	Loss 0.453 (0.610)
Epoch: [25][140/200]	Time 0.502 (0.596)	Data 0.000 (0.072)	Loss 0.682 (0.616)
Epoch: [25][160/200]	Time 0.502 (0.596)	Data 0.000 (0.071)	Loss 0.597 (0.616)
Epoch: [25][180/200]	Time 0.503 (0.596)	Data 0.000 (0.072)	Loss 0.558 (0.619)
Epoch: [25][200/200]	Time 0.524 (0.601)	Data 0.001 (0.077)	Loss 0.676 (0.624)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.217)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.187890768051147
==> Statistics for epoch 26: 600 clusters
Epoch: [26][20/200]	Time 0.503 (0.637)	Data 0.001 (0.112)	Loss 0.680 (0.148)
Epoch: [26][40/200]	Time 0.507 (0.621)	Data 0.002 (0.095)	Loss 0.576 (0.408)
Epoch: [26][60/200]	Time 0.507 (0.613)	Data 0.001 (0.088)	Loss 0.532 (0.487)
Epoch: [26][80/200]	Time 0.509 (0.610)	Data 0.001 (0.085)	Loss 0.598 (0.533)
Epoch: [26][100/200]	Time 0.504 (0.607)	Data 0.001 (0.083)	Loss 0.416 (0.549)
Epoch: [26][120/200]	Time 0.591 (0.606)	Data 0.000 (0.082)	Loss 0.666 (0.567)
Epoch: [26][140/200]	Time 0.505 (0.604)	Data 0.000 (0.079)	Loss 0.612 (0.582)
Epoch: [26][160/200]	Time 0.507 (0.603)	Data 0.000 (0.078)	Loss 0.518 (0.589)
Epoch: [26][180/200]	Time 0.519 (0.602)	Data 0.000 (0.078)	Loss 0.412 (0.597)
Epoch: [26][200/200]	Time 0.505 (0.610)	Data 0.001 (0.085)	Loss 0.573 (0.604)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.220)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.177237510681152
==> Statistics for epoch 27: 602 clusters
Epoch: [27][20/200]	Time 0.507 (0.625)	Data 0.001 (0.102)	Loss 0.692 (0.162)
Epoch: [27][40/200]	Time 0.507 (0.614)	Data 0.001 (0.088)	Loss 0.597 (0.410)
Epoch: [27][60/200]	Time 0.506 (0.606)	Data 0.002 (0.082)	Loss 0.766 (0.491)
Epoch: [27][80/200]	Time 0.508 (0.604)	Data 0.001 (0.081)	Loss 0.558 (0.534)
Epoch: [27][100/200]	Time 0.511 (0.602)	Data 0.001 (0.079)	Loss 0.622 (0.566)
Epoch: [27][120/200]	Time 0.506 (0.600)	Data 0.000 (0.077)	Loss 0.869 (0.581)
Epoch: [27][140/200]	Time 0.598 (0.602)	Data 0.000 (0.078)	Loss 0.545 (0.599)
Epoch: [27][160/200]	Time 0.506 (0.603)	Data 0.000 (0.079)	Loss 0.734 (0.611)
Epoch: [27][180/200]	Time 0.586 (0.603)	Data 0.000 (0.078)	Loss 0.621 (0.619)
Epoch: [27][200/200]	Time 0.518 (0.609)	Data 0.001 (0.084)	Loss 0.432 (0.620)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.290 (0.226)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.03288745880127
==> Statistics for epoch 28: 600 clusters
Epoch: [28][20/200]	Time 0.621 (0.629)	Data 0.001 (0.105)	Loss 0.558 (0.137)
Epoch: [28][40/200]	Time 0.510 (0.605)	Data 0.001 (0.087)	Loss 0.792 (0.374)
Epoch: [28][60/200]	Time 0.499 (0.599)	Data 0.001 (0.079)	Loss 0.687 (0.454)
Epoch: [28][80/200]	Time 0.581 (0.596)	Data 0.001 (0.075)	Loss 0.564 (0.503)
Epoch: [28][100/200]	Time 0.503 (0.594)	Data 0.001 (0.073)	Loss 0.640 (0.535)
Epoch: [28][120/200]	Time 0.500 (0.593)	Data 0.000 (0.072)	Loss 1.000 (0.558)
Epoch: [28][140/200]	Time 0.501 (0.592)	Data 0.000 (0.071)	Loss 0.531 (0.569)
Epoch: [28][160/200]	Time 0.504 (0.591)	Data 0.000 (0.071)	Loss 0.428 (0.579)
Epoch: [28][180/200]	Time 0.581 (0.591)	Data 0.000 (0.071)	Loss 0.654 (0.582)
Epoch: [28][200/200]	Time 0.510 (0.596)	Data 0.001 (0.077)	Loss 0.875 (0.590)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.214)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.834632873535156
==> Statistics for epoch 29: 601 clusters
Epoch: [29][20/200]	Time 0.506 (0.618)	Data 0.001 (0.099)	Loss 0.643 (0.144)
Epoch: [29][40/200]	Time 0.502 (0.605)	Data 0.002 (0.084)	Loss 0.329 (0.380)
Epoch: [29][60/200]	Time 0.509 (0.598)	Data 0.001 (0.079)	Loss 0.646 (0.461)
Epoch: [29][80/200]	Time 0.507 (0.594)	Data 0.001 (0.074)	Loss 0.701 (0.511)
Epoch: [29][100/200]	Time 0.534 (0.592)	Data 0.001 (0.072)	Loss 0.481 (0.529)
Epoch: [29][120/200]	Time 0.499 (0.591)	Data 0.000 (0.071)	Loss 0.487 (0.552)
Epoch: [29][140/200]	Time 0.502 (0.589)	Data 0.000 (0.070)	Loss 0.551 (0.560)
Epoch: [29][160/200]	Time 0.511 (0.590)	Data 0.000 (0.069)	Loss 0.686 (0.571)
Epoch: [29][180/200]	Time 0.501 (0.589)	Data 0.000 (0.069)	Loss 0.543 (0.577)
Epoch: [29][200/200]	Time 0.500 (0.596)	Data 0.001 (0.074)	Loss 0.739 (0.587)
Extract Features: [50/76]	Time 0.180 (0.215)	Data 0.000 (0.022)	
Mean AP: 93.3%

 * Finished epoch  29  model mAP: 93.3%  best: 93.3% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.181 (0.215)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.87151575088501
==> Statistics for epoch 30: 605 clusters
Epoch: [30][20/200]	Time 0.600 (0.619)	Data 0.001 (0.093)	Loss 0.754 (0.158)
Epoch: [30][40/200]	Time 0.517 (0.607)	Data 0.001 (0.082)	Loss 0.479 (0.373)
Epoch: [30][60/200]	Time 0.501 (0.600)	Data 0.001 (0.077)	Loss 0.566 (0.472)
Epoch: [30][80/200]	Time 0.506 (0.597)	Data 0.001 (0.075)	Loss 0.750 (0.520)
Epoch: [30][100/200]	Time 0.506 (0.594)	Data 0.001 (0.073)	Loss 0.609 (0.553)
Epoch: [30][120/200]	Time 0.503 (0.593)	Data 0.000 (0.072)	Loss 0.714 (0.574)
Epoch: [30][140/200]	Time 0.500 (0.592)	Data 0.000 (0.071)	Loss 0.581 (0.592)
Epoch: [30][160/200]	Time 0.502 (0.594)	Data 0.000 (0.072)	Loss 0.784 (0.599)
Epoch: [30][180/200]	Time 0.579 (0.594)	Data 0.000 (0.071)	Loss 0.513 (0.598)
Epoch: [30][200/200]	Time 0.503 (0.599)	Data 0.001 (0.077)	Loss 0.660 (0.607)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.214)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.735856771469116
==> Statistics for epoch 31: 603 clusters
Epoch: [31][20/200]	Time 0.507 (0.623)	Data 0.001 (0.100)	Loss 0.589 (0.159)
Epoch: [31][40/200]	Time 0.516 (0.612)	Data 0.001 (0.088)	Loss 0.786 (0.393)
Epoch: [31][60/200]	Time 0.502 (0.606)	Data 0.001 (0.082)	Loss 0.907 (0.485)
Epoch: [31][80/200]	Time 0.503 (0.603)	Data 0.001 (0.080)	Loss 0.482 (0.514)
Epoch: [31][100/200]	Time 0.501 (0.600)	Data 0.001 (0.077)	Loss 0.561 (0.541)
Epoch: [31][120/200]	Time 0.505 (0.599)	Data 0.000 (0.075)	Loss 1.031 (0.562)
Epoch: [31][140/200]	Time 0.500 (0.599)	Data 0.000 (0.075)	Loss 0.440 (0.572)
Epoch: [31][160/200]	Time 0.502 (0.598)	Data 0.000 (0.074)	Loss 0.664 (0.584)
Epoch: [31][180/200]	Time 0.583 (0.597)	Data 0.000 (0.073)	Loss 0.478 (0.588)
Epoch: [31][200/200]	Time 0.507 (0.606)	Data 0.001 (0.081)	Loss 0.666 (0.594)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.178 (0.216)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.084446668624878
==> Statistics for epoch 32: 604 clusters
Epoch: [32][20/200]	Time 0.513 (0.628)	Data 0.001 (0.102)	Loss 0.668 (0.177)
Epoch: [32][40/200]	Time 0.509 (0.616)	Data 0.001 (0.089)	Loss 0.499 (0.400)
Epoch: [32][60/200]	Time 0.501 (0.611)	Data 0.001 (0.086)	Loss 0.565 (0.487)
Epoch: [32][80/200]	Time 0.503 (0.606)	Data 0.001 (0.081)	Loss 0.435 (0.528)
Epoch: [32][100/200]	Time 0.515 (0.602)	Data 0.001 (0.078)	Loss 0.408 (0.546)
Epoch: [32][120/200]	Time 0.499 (0.600)	Data 0.000 (0.076)	Loss 0.701 (0.567)
Epoch: [32][140/200]	Time 0.579 (0.598)	Data 0.000 (0.075)	Loss 0.801 (0.570)
Epoch: [32][160/200]	Time 0.500 (0.596)	Data 0.000 (0.074)	Loss 0.733 (0.579)
Epoch: [32][180/200]	Time 0.499 (0.595)	Data 0.000 (0.073)	Loss 0.621 (0.582)
Epoch: [32][200/200]	Time 0.498 (0.601)	Data 0.001 (0.079)	Loss 0.772 (0.586)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.217)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.371735095977783
==> Statistics for epoch 33: 605 clusters
Epoch: [33][20/200]	Time 0.509 (0.624)	Data 0.001 (0.103)	Loss 0.653 (0.149)
Epoch: [33][40/200]	Time 0.507 (0.607)	Data 0.001 (0.087)	Loss 0.731 (0.383)
Epoch: [33][60/200]	Time 0.506 (0.601)	Data 0.001 (0.080)	Loss 0.532 (0.457)
Epoch: [33][80/200]	Time 0.596 (0.600)	Data 0.001 (0.080)	Loss 0.634 (0.506)
Epoch: [33][100/200]	Time 0.519 (0.597)	Data 0.001 (0.077)	Loss 0.461 (0.536)
Epoch: [33][120/200]	Time 0.504 (0.598)	Data 0.000 (0.076)	Loss 0.358 (0.565)
Epoch: [33][140/200]	Time 0.578 (0.597)	Data 0.000 (0.075)	Loss 1.006 (0.581)
Epoch: [33][160/200]	Time 0.503 (0.597)	Data 0.000 (0.075)	Loss 0.842 (0.584)
Epoch: [33][180/200]	Time 0.504 (0.596)	Data 0.000 (0.074)	Loss 0.743 (0.591)
Epoch: [33][200/200]	Time 0.501 (0.603)	Data 0.001 (0.081)	Loss 0.597 (0.593)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.215)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.225067615509033
==> Statistics for epoch 34: 605 clusters
Epoch: [34][20/200]	Time 0.616 (0.635)	Data 0.001 (0.111)	Loss 0.691 (0.143)
Epoch: [34][40/200]	Time 0.506 (0.608)	Data 0.001 (0.089)	Loss 0.829 (0.377)
Epoch: [34][60/200]	Time 0.521 (0.602)	Data 0.001 (0.081)	Loss 0.694 (0.464)
Epoch: [34][80/200]	Time 0.587 (0.599)	Data 0.001 (0.077)	Loss 0.506 (0.507)
Epoch: [34][100/200]	Time 0.505 (0.599)	Data 0.001 (0.076)	Loss 0.882 (0.530)
Epoch: [34][120/200]	Time 0.588 (0.599)	Data 0.000 (0.075)	Loss 0.725 (0.549)
Epoch: [34][140/200]	Time 0.604 (0.597)	Data 0.000 (0.073)	Loss 0.626 (0.564)
Epoch: [34][160/200]	Time 0.503 (0.596)	Data 0.000 (0.073)	Loss 0.463 (0.577)
Epoch: [34][180/200]	Time 0.500 (0.595)	Data 0.000 (0.073)	Loss 0.546 (0.584)
Epoch: [34][200/200]	Time 0.511 (0.602)	Data 0.001 (0.079)	Loss 0.872 (0.593)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.216)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.097003698349
==> Statistics for epoch 35: 602 clusters
Epoch: [35][20/200]	Time 0.508 (0.635)	Data 0.001 (0.104)	Loss 0.438 (0.149)
Epoch: [35][40/200]	Time 0.629 (0.621)	Data 0.006 (0.090)	Loss 0.456 (0.371)
Epoch: [35][60/200]	Time 0.507 (0.609)	Data 0.001 (0.081)	Loss 0.515 (0.445)
Epoch: [35][80/200]	Time 0.507 (0.605)	Data 0.001 (0.078)	Loss 0.703 (0.499)
Epoch: [35][100/200]	Time 0.507 (0.600)	Data 0.002 (0.076)	Loss 0.521 (0.523)
Epoch: [35][120/200]	Time 0.589 (0.598)	Data 0.000 (0.074)	Loss 0.553 (0.547)
Epoch: [35][140/200]	Time 0.502 (0.598)	Data 0.000 (0.074)	Loss 0.610 (0.553)
Epoch: [35][160/200]	Time 0.503 (0.597)	Data 0.000 (0.073)	Loss 0.816 (0.567)
Epoch: [35][180/200]	Time 0.501 (0.595)	Data 0.000 (0.072)	Loss 0.702 (0.572)
Epoch: [35][200/200]	Time 0.503 (0.601)	Data 0.001 (0.078)	Loss 0.633 (0.579)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.213)	Data 0.000 (0.020)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.356539487838745
==> Statistics for epoch 36: 601 clusters
Epoch: [36][20/200]	Time 0.503 (0.631)	Data 0.001 (0.108)	Loss 0.666 (0.154)
Epoch: [36][40/200]	Time 0.506 (0.609)	Data 0.001 (0.085)	Loss 0.511 (0.375)
Epoch: [36][60/200]	Time 0.504 (0.599)	Data 0.002 (0.077)	Loss 0.506 (0.473)
Epoch: [36][80/200]	Time 0.509 (0.597)	Data 0.001 (0.075)	Loss 0.473 (0.526)
Epoch: [36][100/200]	Time 0.501 (0.595)	Data 0.001 (0.074)	Loss 0.517 (0.557)
Epoch: [36][120/200]	Time 0.503 (0.593)	Data 0.000 (0.072)	Loss 0.676 (0.578)
Epoch: [36][140/200]	Time 0.503 (0.592)	Data 0.000 (0.071)	Loss 0.623 (0.582)
Epoch: [36][160/200]	Time 0.500 (0.591)	Data 0.000 (0.070)	Loss 0.464 (0.579)
Epoch: [36][180/200]	Time 0.502 (0.590)	Data 0.000 (0.069)	Loss 0.851 (0.587)
Epoch: [36][200/200]	Time 0.501 (0.596)	Data 0.001 (0.074)	Loss 0.478 (0.590)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.178 (0.221)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.328317165374756
==> Statistics for epoch 37: 602 clusters
Epoch: [37][20/200]	Time 0.503 (0.655)	Data 0.001 (0.120)	Loss 0.561 (0.127)
Epoch: [37][40/200]	Time 0.627 (0.623)	Data 0.001 (0.091)	Loss 0.615 (0.349)
Epoch: [37][60/200]	Time 0.503 (0.610)	Data 0.001 (0.083)	Loss 0.539 (0.424)
Epoch: [37][80/200]	Time 0.506 (0.604)	Data 0.001 (0.079)	Loss 0.630 (0.457)
Epoch: [37][100/200]	Time 0.501 (0.601)	Data 0.001 (0.077)	Loss 0.512 (0.487)
Epoch: [37][120/200]	Time 0.581 (0.599)	Data 0.000 (0.075)	Loss 0.544 (0.509)
Epoch: [37][140/200]	Time 0.506 (0.596)	Data 0.000 (0.073)	Loss 0.600 (0.520)
Epoch: [37][160/200]	Time 0.504 (0.596)	Data 0.000 (0.072)	Loss 0.561 (0.534)
Epoch: [37][180/200]	Time 0.501 (0.594)	Data 0.000 (0.070)	Loss 0.712 (0.549)
Epoch: [37][200/200]	Time 0.519 (0.601)	Data 0.001 (0.076)	Loss 0.586 (0.558)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.213)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.455748796463013
==> Statistics for epoch 38: 603 clusters
Epoch: [38][20/200]	Time 0.503 (0.643)	Data 0.001 (0.113)	Loss 0.826 (0.136)
Epoch: [38][40/200]	Time 0.507 (0.630)	Data 0.001 (0.103)	Loss 0.626 (0.362)
Epoch: [38][60/200]	Time 0.502 (0.619)	Data 0.001 (0.095)	Loss 0.572 (0.467)
Epoch: [38][80/200]	Time 0.510 (0.615)	Data 0.001 (0.091)	Loss 0.573 (0.496)
Epoch: [38][100/200]	Time 0.509 (0.612)	Data 0.001 (0.088)	Loss 0.749 (0.527)
Epoch: [38][120/200]	Time 0.589 (0.611)	Data 0.000 (0.086)	Loss 0.814 (0.539)
Epoch: [38][140/200]	Time 0.501 (0.609)	Data 0.000 (0.084)	Loss 0.721 (0.547)
Epoch: [38][160/200]	Time 0.503 (0.608)	Data 0.000 (0.083)	Loss 0.934 (0.553)
Epoch: [38][180/200]	Time 0.506 (0.607)	Data 0.000 (0.083)	Loss 0.391 (0.558)
Epoch: [38][200/200]	Time 0.502 (0.616)	Data 0.001 (0.091)	Loss 0.623 (0.568)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.219)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.225264310836792
==> Statistics for epoch 39: 601 clusters
Epoch: [39][20/200]	Time 0.502 (0.637)	Data 0.001 (0.117)	Loss 0.702 (0.158)
Epoch: [39][40/200]	Time 0.503 (0.618)	Data 0.001 (0.096)	Loss 0.851 (0.369)
Epoch: [39][60/200]	Time 0.500 (0.611)	Data 0.001 (0.090)	Loss 0.509 (0.455)
Epoch: [39][80/200]	Time 0.506 (0.608)	Data 0.001 (0.086)	Loss 0.574 (0.512)
Epoch: [39][100/200]	Time 0.507 (0.606)	Data 0.001 (0.083)	Loss 0.415 (0.534)
Epoch: [39][120/200]	Time 0.506 (0.604)	Data 0.000 (0.082)	Loss 0.798 (0.552)
Epoch: [39][140/200]	Time 0.502 (0.602)	Data 0.000 (0.081)	Loss 0.611 (0.563)
Epoch: [39][160/200]	Time 0.502 (0.601)	Data 0.000 (0.080)	Loss 0.708 (0.565)
Epoch: [39][180/200]	Time 0.504 (0.601)	Data 0.000 (0.079)	Loss 0.780 (0.575)
Epoch: [39][200/200]	Time 0.509 (0.609)	Data 0.001 (0.086)	Loss 0.490 (0.578)
Extract Features: [50/76]	Time 0.180 (0.221)	Data 0.001 (0.025)	
Mean AP: 93.3%

 * Finished epoch  39  model mAP: 93.3%  best: 93.3%

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.222)	Data 0.000 (0.028)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.92641305923462
==> Statistics for epoch 40: 601 clusters
Epoch: [40][20/200]	Time 0.502 (0.643)	Data 0.001 (0.119)	Loss 0.494 (0.130)
Epoch: [40][40/200]	Time 0.502 (0.620)	Data 0.002 (0.096)	Loss 0.596 (0.351)
Epoch: [40][60/200]	Time 0.507 (0.614)	Data 0.001 (0.090)	Loss 0.515 (0.458)
Epoch: [40][80/200]	Time 0.536 (0.611)	Data 0.001 (0.088)	Loss 0.560 (0.501)
Epoch: [40][100/200]	Time 0.504 (0.609)	Data 0.001 (0.086)	Loss 0.644 (0.526)
Epoch: [40][120/200]	Time 0.507 (0.607)	Data 0.000 (0.084)	Loss 0.774 (0.545)
Epoch: [40][140/200]	Time 0.620 (0.607)	Data 0.000 (0.083)	Loss 0.440 (0.558)
Epoch: [40][160/200]	Time 0.501 (0.607)	Data 0.000 (0.083)	Loss 0.755 (0.563)
Epoch: [40][180/200]	Time 0.501 (0.606)	Data 0.000 (0.082)	Loss 0.830 (0.574)
Epoch: [40][200/200]	Time 0.511 (0.614)	Data 0.001 (0.089)	Loss 0.516 (0.577)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.183 (0.225)	Data 0.000 (0.028)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.09749698638916
==> Statistics for epoch 41: 603 clusters
Epoch: [41][20/200]	Time 0.514 (0.635)	Data 0.001 (0.109)	Loss 0.605 (0.151)
Epoch: [41][40/200]	Time 0.509 (0.622)	Data 0.002 (0.097)	Loss 0.953 (0.386)
Epoch: [41][60/200]	Time 0.501 (0.615)	Data 0.001 (0.091)	Loss 0.382 (0.451)
Epoch: [41][80/200]	Time 0.505 (0.610)	Data 0.001 (0.088)	Loss 0.723 (0.493)
Epoch: [41][100/200]	Time 0.504 (0.609)	Data 0.001 (0.085)	Loss 0.695 (0.521)
Epoch: [41][120/200]	Time 0.508 (0.607)	Data 0.000 (0.085)	Loss 0.436 (0.538)
Epoch: [41][140/200]	Time 0.503 (0.607)	Data 0.000 (0.083)	Loss 0.717 (0.542)
Epoch: [41][160/200]	Time 0.504 (0.607)	Data 0.000 (0.083)	Loss 0.572 (0.555)
Epoch: [41][180/200]	Time 0.587 (0.607)	Data 0.000 (0.082)	Loss 0.631 (0.560)
Epoch: [41][200/200]	Time 0.505 (0.614)	Data 0.001 (0.089)	Loss 0.577 (0.568)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.253 (0.221)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.99256443977356
==> Statistics for epoch 42: 600 clusters
Epoch: [42][20/200]	Time 0.605 (0.649)	Data 0.001 (0.119)	Loss 0.359 (0.128)
Epoch: [42][40/200]	Time 0.500 (0.620)	Data 0.001 (0.093)	Loss 0.454 (0.358)
Epoch: [42][60/200]	Time 0.501 (0.610)	Data 0.001 (0.086)	Loss 0.715 (0.454)
Epoch: [42][80/200]	Time 0.538 (0.606)	Data 0.001 (0.083)	Loss 0.635 (0.501)
Epoch: [42][100/200]	Time 0.498 (0.600)	Data 0.001 (0.079)	Loss 0.702 (0.531)
Epoch: [42][120/200]	Time 0.501 (0.599)	Data 0.000 (0.078)	Loss 0.849 (0.540)
Epoch: [42][140/200]	Time 0.501 (0.598)	Data 0.000 (0.075)	Loss 0.680 (0.546)
Epoch: [42][160/200]	Time 0.618 (0.598)	Data 0.000 (0.076)	Loss 0.560 (0.554)
Epoch: [42][180/200]	Time 0.507 (0.598)	Data 0.000 (0.075)	Loss 0.537 (0.562)
Epoch: [42][200/200]	Time 0.502 (0.604)	Data 0.001 (0.081)	Loss 0.667 (0.566)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.181 (0.215)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.81562876701355
==> Statistics for epoch 43: 603 clusters
Epoch: [43][20/200]	Time 0.515 (0.638)	Data 0.001 (0.109)	Loss 0.373 (0.128)
Epoch: [43][40/200]	Time 0.510 (0.619)	Data 0.002 (0.093)	Loss 0.699 (0.376)
Epoch: [43][60/200]	Time 0.513 (0.614)	Data 0.001 (0.088)	Loss 0.739 (0.436)
Epoch: [43][80/200]	Time 0.502 (0.612)	Data 0.001 (0.084)	Loss 0.777 (0.489)
Epoch: [43][100/200]	Time 0.509 (0.607)	Data 0.001 (0.081)	Loss 0.758 (0.520)
Epoch: [43][120/200]	Time 0.502 (0.605)	Data 0.000 (0.079)	Loss 0.616 (0.544)
Epoch: [43][140/200]	Time 0.571 (0.603)	Data 0.000 (0.077)	Loss 0.534 (0.555)
Epoch: [43][160/200]	Time 0.502 (0.602)	Data 0.000 (0.077)	Loss 0.642 (0.564)
Epoch: [43][180/200]	Time 0.499 (0.601)	Data 0.000 (0.077)	Loss 0.561 (0.572)
Epoch: [43][200/200]	Time 0.620 (0.609)	Data 0.001 (0.084)	Loss 0.656 (0.575)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.266 (0.218)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.433276414871216
==> Statistics for epoch 44: 602 clusters
Epoch: [44][20/200]	Time 0.618 (0.638)	Data 0.001 (0.104)	Loss 0.823 (0.154)
Epoch: [44][40/200]	Time 0.503 (0.616)	Data 0.002 (0.094)	Loss 0.682 (0.365)
Epoch: [44][60/200]	Time 0.508 (0.610)	Data 0.001 (0.087)	Loss 0.489 (0.438)
Epoch: [44][80/200]	Time 0.593 (0.608)	Data 0.001 (0.084)	Loss 0.456 (0.483)
Epoch: [44][100/200]	Time 0.508 (0.606)	Data 0.001 (0.082)	Loss 0.555 (0.513)
Epoch: [44][120/200]	Time 0.586 (0.606)	Data 0.000 (0.082)	Loss 0.777 (0.522)
Epoch: [44][140/200]	Time 0.501 (0.604)	Data 0.000 (0.080)	Loss 0.625 (0.535)
Epoch: [44][160/200]	Time 0.501 (0.602)	Data 0.000 (0.079)	Loss 0.546 (0.542)
Epoch: [44][180/200]	Time 0.508 (0.602)	Data 0.000 (0.079)	Loss 0.566 (0.552)
Epoch: [44][200/200]	Time 0.513 (0.609)	Data 0.001 (0.085)	Loss 0.473 (0.558)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.218)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.16767907142639
==> Statistics for epoch 45: 603 clusters
Epoch: [45][20/200]	Time 0.512 (0.633)	Data 0.001 (0.108)	Loss 0.668 (0.143)
Epoch: [45][40/200]	Time 0.500 (0.611)	Data 0.002 (0.087)	Loss 0.721 (0.381)
Epoch: [45][60/200]	Time 0.503 (0.603)	Data 0.001 (0.080)	Loss 0.622 (0.449)
Epoch: [45][80/200]	Time 0.507 (0.599)	Data 0.001 (0.077)	Loss 0.564 (0.481)
Epoch: [45][100/200]	Time 0.519 (0.599)	Data 0.001 (0.076)	Loss 0.683 (0.501)
Epoch: [45][120/200]	Time 0.584 (0.599)	Data 0.000 (0.076)	Loss 0.395 (0.522)
Epoch: [45][140/200]	Time 0.508 (0.598)	Data 0.000 (0.075)	Loss 0.547 (0.532)
Epoch: [45][160/200]	Time 0.500 (0.597)	Data 0.000 (0.075)	Loss 0.753 (0.545)
Epoch: [45][180/200]	Time 0.575 (0.598)	Data 0.000 (0.075)	Loss 0.843 (0.553)
Epoch: [45][200/200]	Time 0.509 (0.604)	Data 0.001 (0.081)	Loss 0.745 (0.564)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.217)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.242416858673096
==> Statistics for epoch 46: 602 clusters
Epoch: [46][20/200]	Time 0.502 (0.615)	Data 0.000 (0.098)	Loss 0.724 (0.140)
Epoch: [46][40/200]	Time 0.504 (0.597)	Data 0.001 (0.079)	Loss 0.526 (0.369)
Epoch: [46][60/200]	Time 0.507 (0.597)	Data 0.001 (0.077)	Loss 0.550 (0.453)
Epoch: [46][80/200]	Time 0.542 (0.597)	Data 0.000 (0.076)	Loss 0.807 (0.493)
Epoch: [46][100/200]	Time 0.509 (0.598)	Data 0.000 (0.075)	Loss 0.575 (0.526)
Epoch: [46][120/200]	Time 0.585 (0.599)	Data 0.000 (0.075)	Loss 0.616 (0.538)
Epoch: [46][140/200]	Time 0.503 (0.599)	Data 0.000 (0.075)	Loss 0.698 (0.554)
Epoch: [46][160/200]	Time 0.501 (0.597)	Data 0.000 (0.074)	Loss 0.634 (0.567)
Epoch: [46][180/200]	Time 0.503 (0.596)	Data 0.000 (0.074)	Loss 0.669 (0.569)
Epoch: [46][200/200]	Time 0.503 (0.603)	Data 0.000 (0.079)	Loss 0.554 (0.571)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.214)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.001864910125732
==> Statistics for epoch 47: 602 clusters
Epoch: [47][20/200]	Time 0.518 (0.636)	Data 0.001 (0.105)	Loss 0.604 (0.133)
Epoch: [47][40/200]	Time 0.614 (0.612)	Data 0.001 (0.083)	Loss 0.600 (0.354)
Epoch: [47][60/200]	Time 0.502 (0.605)	Data 0.001 (0.078)	Loss 0.587 (0.447)
Epoch: [47][80/200]	Time 0.508 (0.602)	Data 0.001 (0.077)	Loss 0.502 (0.482)
Epoch: [47][100/200]	Time 0.584 (0.602)	Data 0.001 (0.077)	Loss 0.511 (0.510)
Epoch: [47][120/200]	Time 0.503 (0.598)	Data 0.000 (0.074)	Loss 0.724 (0.528)
Epoch: [47][140/200]	Time 0.498 (0.596)	Data 0.000 (0.073)	Loss 0.553 (0.537)
Epoch: [47][160/200]	Time 0.504 (0.597)	Data 0.000 (0.073)	Loss 0.634 (0.552)
Epoch: [47][180/200]	Time 0.514 (0.598)	Data 0.000 (0.073)	Loss 0.645 (0.555)
Epoch: [47][200/200]	Time 0.505 (0.605)	Data 0.001 (0.080)	Loss 0.605 (0.562)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.179 (0.219)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.68511152267456
==> Statistics for epoch 48: 603 clusters
Epoch: [48][20/200]	Time 0.509 (0.626)	Data 0.001 (0.105)	Loss 0.637 (0.140)
Epoch: [48][40/200]	Time 0.503 (0.603)	Data 0.001 (0.084)	Loss 0.481 (0.362)
Epoch: [48][60/200]	Time 0.504 (0.602)	Data 0.001 (0.081)	Loss 0.387 (0.440)
Epoch: [48][80/200]	Time 0.504 (0.599)	Data 0.001 (0.078)	Loss 0.500 (0.486)
Epoch: [48][100/200]	Time 0.505 (0.599)	Data 0.001 (0.076)	Loss 0.659 (0.511)
Epoch: [48][120/200]	Time 0.579 (0.597)	Data 0.000 (0.074)	Loss 0.483 (0.530)
Epoch: [48][140/200]	Time 0.612 (0.596)	Data 0.000 (0.073)	Loss 0.522 (0.543)
Epoch: [48][160/200]	Time 0.517 (0.594)	Data 0.000 (0.072)	Loss 0.488 (0.547)
Epoch: [48][180/200]	Time 0.517 (0.596)	Data 0.004 (0.074)	Loss 0.565 (0.553)
Epoch: [48][200/200]	Time 0.505 (0.602)	Data 0.001 (0.080)	Loss 0.660 (0.559)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.182 (0.217)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.408421277999878
==> Statistics for epoch 49: 603 clusters
Epoch: [49][20/200]	Time 0.510 (0.634)	Data 0.001 (0.108)	Loss 0.514 (0.128)
Epoch: [49][40/200]	Time 0.509 (0.615)	Data 0.002 (0.090)	Loss 0.751 (0.341)
Epoch: [49][60/200]	Time 0.501 (0.606)	Data 0.001 (0.082)	Loss 0.822 (0.419)
Epoch: [49][80/200]	Time 0.508 (0.605)	Data 0.001 (0.082)	Loss 0.598 (0.475)
Epoch: [49][100/200]	Time 0.614 (0.604)	Data 0.002 (0.080)	Loss 0.727 (0.501)
Epoch: [49][120/200]	Time 0.504 (0.603)	Data 0.000 (0.080)	Loss 0.692 (0.520)
Epoch: [49][140/200]	Time 0.598 (0.602)	Data 0.000 (0.079)	Loss 0.381 (0.528)
Epoch: [49][160/200]	Time 0.506 (0.603)	Data 0.000 (0.079)	Loss 0.660 (0.542)
Epoch: [49][180/200]	Time 0.504 (0.603)	Data 0.000 (0.079)	Loss 0.703 (0.549)
Epoch: [49][200/200]	Time 0.514 (0.610)	Data 0.000 (0.086)	Loss 0.695 (0.559)
Extract Features: [50/76]	Time 0.272 (0.223)	Data 0.000 (0.026)	
Mean AP: 93.3%

 * Finished epoch  49  model mAP: 93.3%  best: 93.3% *

==> Test with the best model:
=> Loaded checkpoint 'log/market/resnet152_cion/model_best.pth.tar'
Extract Features: [50/76]	Time 0.182 (0.220)	Data 0.000 (0.025)	
Mean AP: 93.3%
CMC Scores:
  top-1          97.1%
  top-5          99.0%
  top-10         99.4%
Total running time:  2:11:38.005943
