==========
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='resnet_ibn101a', pretrained_path='/home/ma-user/work/Projects/ReIDNet_Finetune/TransReID/log/bs_exp/ResNet101_IBN_MSMT17/64bs_lr0.0004_ep120_warm20_seed0/resnet101_ibn_120.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/msmt2market/resnet101_ibn_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
checkpoint loaded!
optimizer: SGD
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.145 (0.542)	Data 0.000 (0.018)	
Computing jaccard distance...
Jaccard distance computing time cost: 21.7599937915802
Clustering criterion: eps: 0.600
==> Statistics for epoch 0: 554 clusters
Epoch: [0][20/200]	Time 0.408 (0.881)	Data 0.001 (0.099)	Loss 3.009 (3.153)
Epoch: [0][40/200]	Time 0.411 (0.681)	Data 0.001 (0.078)	Loss 2.690 (3.021)
Epoch: [0][60/200]	Time 0.405 (0.616)	Data 0.001 (0.072)	Loss 2.082 (2.787)
Epoch: [0][80/200]	Time 0.404 (0.583)	Data 0.000 (0.070)	Loss 2.076 (2.603)
Epoch: [0][100/200]	Time 0.408 (0.563)	Data 0.000 (0.069)	Loss 2.005 (2.480)
Epoch: [0][120/200]	Time 1.735 (0.562)	Data 1.270 (0.080)	Loss 1.484 (2.364)
Epoch: [0][140/200]	Time 0.406 (0.552)	Data 0.001 (0.077)	Loss 2.064 (2.278)
Epoch: [0][160/200]	Time 0.411 (0.543)	Data 0.001 (0.075)	Loss 1.808 (2.203)
Epoch: [0][180/200]	Time 0.424 (0.536)	Data 0.000 (0.074)	Loss 1.639 (2.152)
Epoch: [0][200/200]	Time 0.408 (0.532)	Data 0.000 (0.073)	Loss 1.665 (2.105)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.182)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.803143739700317
==> Statistics for epoch 1: 594 clusters
Epoch: [1][20/200]	Time 0.410 (0.539)	Data 0.001 (0.113)	Loss 1.833 (0.419)
Epoch: [1][40/200]	Time 0.408 (0.517)	Data 0.001 (0.092)	Loss 1.743 (0.992)
Epoch: [1][60/200]	Time 0.542 (0.508)	Data 0.001 (0.086)	Loss 1.330 (1.184)
Epoch: [1][80/200]	Time 0.398 (0.505)	Data 0.001 (0.083)	Loss 1.419 (1.283)
Epoch: [1][100/200]	Time 0.399 (0.504)	Data 0.001 (0.080)	Loss 1.689 (1.334)
Epoch: [1][120/200]	Time 0.408 (0.499)	Data 0.000 (0.078)	Loss 1.663 (1.356)
Epoch: [1][140/200]	Time 0.409 (0.497)	Data 0.000 (0.077)	Loss 1.885 (1.376)
Epoch: [1][160/200]	Time 0.410 (0.497)	Data 0.000 (0.076)	Loss 2.192 (1.393)
Epoch: [1][180/200]	Time 0.404 (0.496)	Data 0.000 (0.075)	Loss 1.840 (1.406)
Epoch: [1][200/200]	Time 0.426 (0.503)	Data 0.001 (0.082)	Loss 1.705 (1.410)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.191)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.322934865951538
==> Statistics for epoch 2: 612 clusters
Epoch: [2][20/200]	Time 1.876 (0.544)	Data 1.431 (0.111)	Loss 1.359 (0.294)
Epoch: [2][40/200]	Time 0.412 (0.524)	Data 0.001 (0.093)	Loss 1.258 (0.833)
Epoch: [2][60/200]	Time 0.441 (0.509)	Data 0.001 (0.083)	Loss 1.306 (1.040)
Epoch: [2][80/200]	Time 0.413 (0.509)	Data 0.004 (0.082)	Loss 1.332 (1.140)
Epoch: [2][100/200]	Time 0.406 (0.506)	Data 0.001 (0.081)	Loss 1.910 (1.189)
Epoch: [2][120/200]	Time 0.395 (0.503)	Data 0.001 (0.079)	Loss 1.017 (1.227)
Epoch: [2][140/200]	Time 0.411 (0.502)	Data 0.001 (0.077)	Loss 1.276 (1.251)
Epoch: [2][160/200]	Time 0.532 (0.501)	Data 0.001 (0.075)	Loss 1.748 (1.259)
Epoch: [2][180/200]	Time 0.415 (0.499)	Data 0.001 (0.075)	Loss 1.330 (1.263)
Epoch: [2][200/200]	Time 0.399 (0.498)	Data 0.001 (0.074)	Loss 1.199 (1.262)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.184)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.2952561378479
==> Statistics for epoch 3: 615 clusters
Epoch: [3][20/200]	Time 1.741 (0.528)	Data 1.285 (0.099)	Loss 1.101 (0.241)
Epoch: [3][40/200]	Time 0.415 (0.508)	Data 0.001 (0.082)	Loss 0.882 (0.725)
Epoch: [3][60/200]	Time 0.421 (0.499)	Data 0.001 (0.076)	Loss 1.119 (0.923)
Epoch: [3][80/200]	Time 0.405 (0.497)	Data 0.001 (0.073)	Loss 1.504 (1.006)
Epoch: [3][100/200]	Time 0.412 (0.497)	Data 0.001 (0.072)	Loss 1.344 (1.054)
Epoch: [3][120/200]	Time 0.519 (0.495)	Data 0.001 (0.070)	Loss 1.077 (1.085)
Epoch: [3][140/200]	Time 0.405 (0.494)	Data 0.001 (0.069)	Loss 1.087 (1.108)
Epoch: [3][160/200]	Time 0.412 (0.493)	Data 0.001 (0.069)	Loss 1.193 (1.124)
Epoch: [3][180/200]	Time 0.514 (0.492)	Data 0.001 (0.068)	Loss 1.429 (1.137)
Epoch: [3][200/200]	Time 0.416 (0.491)	Data 0.001 (0.068)	Loss 1.331 (1.150)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.145 (0.180)	Data 0.000 (0.019)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.509411096572876
==> Statistics for epoch 4: 616 clusters
Epoch: [4][20/200]	Time 1.682 (0.521)	Data 1.241 (0.096)	Loss 1.059 (0.226)
Epoch: [4][40/200]	Time 0.404 (0.502)	Data 0.001 (0.080)	Loss 1.342 (0.712)
Epoch: [4][60/200]	Time 0.414 (0.492)	Data 0.001 (0.073)	Loss 1.104 (0.874)
Epoch: [4][80/200]	Time 0.503 (0.489)	Data 0.001 (0.069)	Loss 1.279 (0.952)
Epoch: [4][100/200]	Time 0.398 (0.488)	Data 0.001 (0.068)	Loss 1.731 (1.008)
Epoch: [4][120/200]	Time 0.421 (0.487)	Data 0.001 (0.067)	Loss 1.461 (1.037)
Epoch: [4][140/200]	Time 0.418 (0.487)	Data 0.001 (0.067)	Loss 0.886 (1.061)
Epoch: [4][160/200]	Time 0.421 (0.487)	Data 0.001 (0.066)	Loss 1.261 (1.083)
Epoch: [4][180/200]	Time 0.397 (0.487)	Data 0.001 (0.066)	Loss 1.174 (1.095)
Epoch: [4][200/200]	Time 0.409 (0.487)	Data 0.001 (0.066)	Loss 1.538 (1.105)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.146 (0.178)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.150617599487305
==> Statistics for epoch 5: 621 clusters
Epoch: [5][20/200]	Time 1.762 (0.526)	Data 1.297 (0.102)	Loss 1.114 (0.210)
Epoch: [5][40/200]	Time 0.412 (0.507)	Data 0.001 (0.085)	Loss 1.154 (0.655)
Epoch: [5][60/200]	Time 0.507 (0.501)	Data 0.001 (0.078)	Loss 1.608 (0.812)
Epoch: [5][80/200]	Time 0.408 (0.496)	Data 0.001 (0.074)	Loss 1.435 (0.889)
Epoch: [5][100/200]	Time 0.447 (0.495)	Data 0.001 (0.072)	Loss 1.321 (0.949)
Epoch: [5][120/200]	Time 0.413 (0.494)	Data 0.001 (0.072)	Loss 0.900 (0.976)
Epoch: [5][140/200]	Time 0.403 (0.493)	Data 0.001 (0.071)	Loss 1.102 (0.996)
Epoch: [5][160/200]	Time 0.397 (0.492)	Data 0.001 (0.070)	Loss 0.918 (1.009)
Epoch: [5][180/200]	Time 0.414 (0.492)	Data 0.001 (0.070)	Loss 0.948 (1.016)
Epoch: [5][200/200]	Time 0.403 (0.491)	Data 0.001 (0.069)	Loss 0.916 (1.025)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.179)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.6589572429657
==> Statistics for epoch 6: 618 clusters
Epoch: [6][20/200]	Time 1.664 (0.517)	Data 1.209 (0.094)	Loss 1.072 (0.212)
Epoch: [6][40/200]	Time 0.410 (0.502)	Data 0.001 (0.077)	Loss 0.928 (0.634)
Epoch: [6][60/200]	Time 0.401 (0.495)	Data 0.001 (0.072)	Loss 1.106 (0.757)
Epoch: [6][80/200]	Time 0.535 (0.493)	Data 0.001 (0.072)	Loss 0.649 (0.835)
Epoch: [6][100/200]	Time 0.400 (0.490)	Data 0.001 (0.070)	Loss 1.273 (0.890)
Epoch: [6][120/200]	Time 0.420 (0.489)	Data 0.001 (0.068)	Loss 1.266 (0.928)
Epoch: [6][140/200]	Time 0.396 (0.488)	Data 0.001 (0.068)	Loss 1.019 (0.946)
Epoch: [6][160/200]	Time 0.408 (0.488)	Data 0.001 (0.067)	Loss 0.810 (0.960)
Epoch: [6][180/200]	Time 0.413 (0.488)	Data 0.001 (0.067)	Loss 0.851 (0.978)
Epoch: [6][200/200]	Time 0.409 (0.488)	Data 0.001 (0.066)	Loss 0.789 (0.981)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.186)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.56112504005432
==> Statistics for epoch 7: 615 clusters
Epoch: [7][20/200]	Time 1.656 (0.515)	Data 1.193 (0.094)	Loss 1.136 (0.214)
Epoch: [7][40/200]	Time 0.417 (0.501)	Data 0.001 (0.077)	Loss 1.002 (0.631)
Epoch: [7][60/200]	Time 0.404 (0.495)	Data 0.001 (0.071)	Loss 1.202 (0.771)
Epoch: [7][80/200]	Time 0.410 (0.493)	Data 0.001 (0.068)	Loss 1.179 (0.834)
Epoch: [7][100/200]	Time 0.420 (0.491)	Data 0.001 (0.067)	Loss 0.896 (0.866)
Epoch: [7][120/200]	Time 0.415 (0.489)	Data 0.001 (0.066)	Loss 0.977 (0.889)
Epoch: [7][140/200]	Time 0.412 (0.491)	Data 0.001 (0.068)	Loss 0.874 (0.905)
Epoch: [7][160/200]	Time 0.492 (0.491)	Data 0.001 (0.068)	Loss 0.731 (0.920)
Epoch: [7][180/200]	Time 0.414 (0.491)	Data 0.001 (0.067)	Loss 0.817 (0.933)
Epoch: [7][200/200]	Time 0.403 (0.490)	Data 0.001 (0.067)	Loss 0.955 (0.944)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.181)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.367083072662354
==> Statistics for epoch 8: 615 clusters
Epoch: [8][20/200]	Time 1.700 (0.525)	Data 1.235 (0.101)	Loss 0.806 (0.164)
Epoch: [8][40/200]	Time 0.436 (0.505)	Data 0.001 (0.083)	Loss 0.804 (0.542)
Epoch: [8][60/200]	Time 0.405 (0.498)	Data 0.001 (0.075)	Loss 0.805 (0.649)
Epoch: [8][80/200]	Time 0.415 (0.492)	Data 0.001 (0.072)	Loss 0.968 (0.726)
Epoch: [8][100/200]	Time 0.415 (0.490)	Data 0.001 (0.070)	Loss 0.923 (0.775)
Epoch: [8][120/200]	Time 0.413 (0.488)	Data 0.001 (0.069)	Loss 0.869 (0.799)
Epoch: [8][140/200]	Time 0.408 (0.487)	Data 0.001 (0.068)	Loss 0.963 (0.823)
Epoch: [8][160/200]	Time 0.401 (0.487)	Data 0.001 (0.067)	Loss 0.718 (0.848)
Epoch: [8][180/200]	Time 0.506 (0.487)	Data 0.001 (0.067)	Loss 1.015 (0.857)
Epoch: [8][200/200]	Time 0.403 (0.487)	Data 0.001 (0.067)	Loss 1.163 (0.867)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.181)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.413299560546875
==> Statistics for epoch 9: 613 clusters
Epoch: [9][20/200]	Time 1.784 (0.524)	Data 1.339 (0.107)	Loss 1.046 (0.195)
Epoch: [9][40/200]	Time 0.421 (0.504)	Data 0.001 (0.085)	Loss 0.944 (0.578)
Epoch: [9][60/200]	Time 0.410 (0.499)	Data 0.001 (0.079)	Loss 0.887 (0.691)
Epoch: [9][80/200]	Time 0.503 (0.496)	Data 0.001 (0.074)	Loss 0.946 (0.752)
Epoch: [9][100/200]	Time 0.412 (0.492)	Data 0.001 (0.072)	Loss 0.844 (0.795)
Epoch: [9][120/200]	Time 0.414 (0.491)	Data 0.001 (0.070)	Loss 0.800 (0.824)
Epoch: [9][140/200]	Time 0.525 (0.491)	Data 0.001 (0.069)	Loss 1.027 (0.849)
Epoch: [9][160/200]	Time 0.415 (0.490)	Data 0.001 (0.069)	Loss 0.889 (0.857)
Epoch: [9][180/200]	Time 0.402 (0.490)	Data 0.001 (0.069)	Loss 0.776 (0.869)
Epoch: [9][200/200]	Time 0.413 (0.489)	Data 0.001 (0.068)	Loss 1.032 (0.876)
Extract Features: [50/76]	Time 0.143 (0.180)	Data 0.000 (0.020)	
Mean AP: 93.2%

 * Finished epoch   9  model mAP: 93.2%  best: 93.2% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.182)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.194966316223145
==> Statistics for epoch 10: 614 clusters
Epoch: [10][20/200]	Time 1.843 (0.554)	Data 1.371 (0.111)	Loss 1.110 (0.172)
Epoch: [10][40/200]	Time 0.442 (0.521)	Data 0.001 (0.087)	Loss 0.928 (0.512)
Epoch: [10][60/200]	Time 0.409 (0.509)	Data 0.001 (0.079)	Loss 0.960 (0.650)
Epoch: [10][80/200]	Time 0.415 (0.502)	Data 0.001 (0.075)	Loss 0.794 (0.716)
Epoch: [10][100/200]	Time 0.418 (0.500)	Data 0.001 (0.073)	Loss 0.851 (0.737)
Epoch: [10][120/200]	Time 0.417 (0.499)	Data 0.001 (0.073)	Loss 1.107 (0.770)
Epoch: [10][140/200]	Time 0.402 (0.498)	Data 0.001 (0.072)	Loss 0.895 (0.793)
Epoch: [10][160/200]	Time 0.401 (0.498)	Data 0.001 (0.072)	Loss 0.904 (0.810)
Epoch: [10][180/200]	Time 0.411 (0.497)	Data 0.001 (0.071)	Loss 1.123 (0.810)
Epoch: [10][200/200]	Time 0.482 (0.496)	Data 0.001 (0.071)	Loss 0.913 (0.816)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.180)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.014808416366577
==> Statistics for epoch 11: 613 clusters
Epoch: [11][20/200]	Time 1.983 (0.548)	Data 1.546 (0.118)	Loss 1.130 (0.180)
Epoch: [11][40/200]	Time 0.540 (0.523)	Data 0.001 (0.096)	Loss 0.906 (0.525)
Epoch: [11][60/200]	Time 0.409 (0.512)	Data 0.001 (0.086)	Loss 1.349 (0.642)
Epoch: [11][80/200]	Time 0.412 (0.506)	Data 0.001 (0.081)	Loss 0.774 (0.715)
Epoch: [11][100/200]	Time 0.412 (0.506)	Data 0.005 (0.081)	Loss 0.756 (0.751)
Epoch: [11][120/200]	Time 0.400 (0.504)	Data 0.001 (0.080)	Loss 0.853 (0.762)
Epoch: [11][140/200]	Time 0.408 (0.502)	Data 0.001 (0.078)	Loss 1.107 (0.780)
Epoch: [11][160/200]	Time 0.411 (0.502)	Data 0.001 (0.077)	Loss 0.594 (0.785)
Epoch: [11][180/200]	Time 0.546 (0.500)	Data 0.002 (0.076)	Loss 0.981 (0.788)
Epoch: [11][200/200]	Time 0.405 (0.499)	Data 0.001 (0.076)	Loss 0.938 (0.801)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.188)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.535086393356323
==> Statistics for epoch 12: 614 clusters
Epoch: [12][20/200]	Time 1.791 (0.527)	Data 1.320 (0.103)	Loss 0.783 (0.149)
Epoch: [12][40/200]	Time 0.421 (0.505)	Data 0.001 (0.084)	Loss 1.097 (0.512)
Epoch: [12][60/200]	Time 0.448 (0.502)	Data 0.001 (0.079)	Loss 1.116 (0.636)
Epoch: [12][80/200]	Time 0.419 (0.498)	Data 0.001 (0.075)	Loss 0.826 (0.694)
Epoch: [12][100/200]	Time 0.425 (0.495)	Data 0.001 (0.073)	Loss 0.881 (0.722)
Epoch: [12][120/200]	Time 0.396 (0.495)	Data 0.001 (0.072)	Loss 1.046 (0.739)
Epoch: [12][140/200]	Time 0.505 (0.494)	Data 0.001 (0.071)	Loss 0.960 (0.745)
Epoch: [12][160/200]	Time 0.413 (0.494)	Data 0.001 (0.071)	Loss 0.775 (0.755)
Epoch: [12][180/200]	Time 0.395 (0.493)	Data 0.001 (0.070)	Loss 0.671 (0.762)
Epoch: [12][200/200]	Time 0.495 (0.492)	Data 0.001 (0.070)	Loss 0.764 (0.774)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.177)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.46575951576233
==> Statistics for epoch 13: 617 clusters
Epoch: [13][20/200]	Time 1.780 (0.534)	Data 1.294 (0.108)	Loss 0.803 (0.160)
Epoch: [13][40/200]	Time 0.440 (0.508)	Data 0.001 (0.084)	Loss 1.156 (0.517)
Epoch: [13][60/200]	Time 0.412 (0.502)	Data 0.001 (0.079)	Loss 1.090 (0.637)
Epoch: [13][80/200]	Time 0.524 (0.500)	Data 0.001 (0.076)	Loss 0.515 (0.687)
Epoch: [13][100/200]	Time 0.412 (0.496)	Data 0.001 (0.074)	Loss 1.202 (0.716)
Epoch: [13][120/200]	Time 0.414 (0.496)	Data 0.001 (0.073)	Loss 0.691 (0.732)
Epoch: [13][140/200]	Time 0.520 (0.497)	Data 0.001 (0.073)	Loss 0.809 (0.740)
Epoch: [13][160/200]	Time 0.413 (0.495)	Data 0.001 (0.072)	Loss 0.633 (0.749)
Epoch: [13][180/200]	Time 0.408 (0.495)	Data 0.002 (0.072)	Loss 0.761 (0.755)
Epoch: [13][200/200]	Time 0.417 (0.494)	Data 0.001 (0.071)	Loss 0.693 (0.761)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.183)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.067322969436646
==> Statistics for epoch 14: 615 clusters
Epoch: [14][20/200]	Time 1.965 (0.546)	Data 1.497 (0.111)	Loss 0.725 (0.142)
Epoch: [14][40/200]	Time 0.409 (0.523)	Data 0.001 (0.095)	Loss 0.847 (0.481)
Epoch: [14][60/200]	Time 0.433 (0.516)	Data 0.001 (0.089)	Loss 0.893 (0.591)
Epoch: [14][80/200]	Time 0.405 (0.514)	Data 0.002 (0.086)	Loss 0.749 (0.640)
Epoch: [14][100/200]	Time 0.419 (0.508)	Data 0.001 (0.082)	Loss 0.591 (0.669)
Epoch: [14][120/200]	Time 0.414 (0.505)	Data 0.001 (0.079)	Loss 0.716 (0.693)
Epoch: [14][140/200]	Time 0.410 (0.502)	Data 0.001 (0.077)	Loss 0.886 (0.707)
Epoch: [14][160/200]	Time 0.410 (0.500)	Data 0.001 (0.075)	Loss 0.791 (0.715)
Epoch: [14][180/200]	Time 0.391 (0.500)	Data 0.001 (0.075)	Loss 1.021 (0.725)
Epoch: [14][200/200]	Time 0.400 (0.499)	Data 0.001 (0.075)	Loss 0.845 (0.735)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.152 (0.187)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.49334955215454
==> Statistics for epoch 15: 615 clusters
Epoch: [15][20/200]	Time 1.761 (0.531)	Data 1.286 (0.104)	Loss 0.589 (0.131)
Epoch: [15][40/200]	Time 0.439 (0.512)	Data 0.001 (0.087)	Loss 0.856 (0.447)
Epoch: [15][60/200]	Time 0.399 (0.501)	Data 0.001 (0.079)	Loss 0.785 (0.549)
Epoch: [15][80/200]	Time 0.407 (0.497)	Data 0.001 (0.075)	Loss 1.104 (0.612)
Epoch: [15][100/200]	Time 0.399 (0.498)	Data 0.001 (0.075)	Loss 0.702 (0.643)
Epoch: [15][120/200]	Time 0.534 (0.496)	Data 0.001 (0.074)	Loss 1.174 (0.667)
Epoch: [15][140/200]	Time 0.409 (0.496)	Data 0.001 (0.074)	Loss 1.254 (0.679)
Epoch: [15][160/200]	Time 0.414 (0.495)	Data 0.001 (0.074)	Loss 1.061 (0.688)
Epoch: [15][180/200]	Time 0.488 (0.495)	Data 0.001 (0.074)	Loss 0.828 (0.693)
Epoch: [15][200/200]	Time 0.423 (0.494)	Data 0.005 (0.072)	Loss 0.673 (0.696)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.193)	Data 0.000 (0.033)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.04259967803955
==> Statistics for epoch 16: 616 clusters
Epoch: [16][20/200]	Time 1.900 (0.546)	Data 1.416 (0.117)	Loss 0.661 (0.121)
Epoch: [16][40/200]	Time 0.415 (0.522)	Data 0.001 (0.098)	Loss 0.747 (0.407)
Epoch: [16][60/200]	Time 0.413 (0.513)	Data 0.001 (0.090)	Loss 0.757 (0.528)
Epoch: [16][80/200]	Time 0.396 (0.508)	Data 0.002 (0.086)	Loss 1.014 (0.587)
Epoch: [16][100/200]	Time 0.396 (0.506)	Data 0.001 (0.083)	Loss 0.685 (0.618)
Epoch: [16][120/200]	Time 0.400 (0.506)	Data 0.001 (0.083)	Loss 0.697 (0.644)
Epoch: [16][140/200]	Time 0.415 (0.503)	Data 0.003 (0.080)	Loss 0.879 (0.661)
Epoch: [16][160/200]	Time 0.515 (0.505)	Data 0.001 (0.082)	Loss 0.680 (0.677)
Epoch: [16][180/200]	Time 0.404 (0.505)	Data 0.001 (0.081)	Loss 0.495 (0.684)
Epoch: [16][200/200]	Time 0.413 (0.503)	Data 0.001 (0.080)	Loss 0.763 (0.691)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.146 (0.187)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.47752285003662
==> Statistics for epoch 17: 613 clusters
Epoch: [17][20/200]	Time 1.715 (0.525)	Data 1.229 (0.097)	Loss 0.501 (0.117)
Epoch: [17][40/200]	Time 0.430 (0.511)	Data 0.001 (0.086)	Loss 0.622 (0.401)
Epoch: [17][60/200]	Time 0.395 (0.506)	Data 0.001 (0.082)	Loss 0.649 (0.504)
Epoch: [17][80/200]	Time 0.516 (0.501)	Data 0.001 (0.077)	Loss 0.719 (0.555)
Epoch: [17][100/200]	Time 0.399 (0.497)	Data 0.001 (0.076)	Loss 0.632 (0.588)
Epoch: [17][120/200]	Time 0.401 (0.497)	Data 0.001 (0.074)	Loss 0.543 (0.615)
Epoch: [17][140/200]	Time 0.503 (0.495)	Data 0.001 (0.073)	Loss 0.703 (0.632)
Epoch: [17][160/200]	Time 0.409 (0.493)	Data 0.001 (0.072)	Loss 0.627 (0.640)
Epoch: [17][180/200]	Time 0.401 (0.493)	Data 0.001 (0.072)	Loss 0.745 (0.648)
Epoch: [17][200/200]	Time 0.416 (0.492)	Data 0.001 (0.071)	Loss 0.536 (0.655)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.183)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.260871648788452
==> Statistics for epoch 18: 616 clusters
Epoch: [18][20/200]	Time 1.949 (0.541)	Data 1.496 (0.110)	Loss 0.835 (0.135)
Epoch: [18][40/200]	Time 0.423 (0.517)	Data 0.001 (0.091)	Loss 0.729 (0.422)
Epoch: [18][60/200]	Time 0.400 (0.511)	Data 0.001 (0.085)	Loss 0.570 (0.529)
Epoch: [18][80/200]	Time 0.417 (0.509)	Data 0.001 (0.083)	Loss 0.623 (0.571)
Epoch: [18][100/200]	Time 0.566 (0.505)	Data 0.001 (0.081)	Loss 0.580 (0.593)
Epoch: [18][120/200]	Time 0.394 (0.502)	Data 0.001 (0.078)	Loss 0.856 (0.620)
Epoch: [18][140/200]	Time 0.414 (0.501)	Data 0.001 (0.076)	Loss 0.872 (0.634)
Epoch: [18][160/200]	Time 0.408 (0.499)	Data 0.001 (0.075)	Loss 0.569 (0.643)
Epoch: [18][180/200]	Time 0.406 (0.498)	Data 0.001 (0.075)	Loss 0.785 (0.647)
Epoch: [18][200/200]	Time 0.410 (0.497)	Data 0.001 (0.074)	Loss 0.559 (0.653)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.181)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.27554965019226
==> Statistics for epoch 19: 613 clusters
Epoch: [19][20/200]	Time 1.795 (0.525)	Data 1.340 (0.101)	Loss 0.597 (0.120)
Epoch: [19][40/200]	Time 0.428 (0.500)	Data 0.001 (0.082)	Loss 0.604 (0.383)
Epoch: [19][60/200]	Time 0.417 (0.496)	Data 0.001 (0.078)	Loss 0.791 (0.475)
Epoch: [19][80/200]	Time 0.411 (0.494)	Data 0.001 (0.075)	Loss 0.770 (0.531)
Epoch: [19][100/200]	Time 0.413 (0.493)	Data 0.001 (0.074)	Loss 0.571 (0.575)
Epoch: [19][120/200]	Time 0.398 (0.493)	Data 0.001 (0.072)	Loss 0.717 (0.591)
Epoch: [19][140/200]	Time 0.550 (0.494)	Data 0.002 (0.073)	Loss 0.763 (0.600)
Epoch: [19][160/200]	Time 0.410 (0.494)	Data 0.001 (0.074)	Loss 0.654 (0.616)
Epoch: [19][180/200]	Time 0.411 (0.495)	Data 0.001 (0.074)	Loss 0.741 (0.627)
Epoch: [19][200/200]	Time 0.402 (0.495)	Data 0.001 (0.075)	Loss 0.598 (0.628)
Extract Features: [50/76]	Time 0.142 (0.185)	Data 0.000 (0.025)	
Mean AP: 93.4%

 * Finished epoch  19  model mAP: 93.4%  best: 93.4% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.185)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.47775936126709
==> Statistics for epoch 20: 614 clusters
Epoch: [20][20/200]	Time 1.957 (0.541)	Data 1.487 (0.116)	Loss 0.558 (0.108)
Epoch: [20][40/200]	Time 0.395 (0.515)	Data 0.001 (0.092)	Loss 0.791 (0.396)
Epoch: [20][60/200]	Time 0.414 (0.503)	Data 0.001 (0.082)	Loss 0.612 (0.490)
Epoch: [20][80/200]	Time 0.410 (0.498)	Data 0.001 (0.078)	Loss 0.623 (0.529)
Epoch: [20][100/200]	Time 0.409 (0.495)	Data 0.001 (0.075)	Loss 0.598 (0.558)
Epoch: [20][120/200]	Time 0.418 (0.494)	Data 0.000 (0.074)	Loss 0.466 (0.568)
Epoch: [20][140/200]	Time 0.411 (0.494)	Data 0.001 (0.073)	Loss 0.761 (0.582)
Epoch: [20][160/200]	Time 0.406 (0.494)	Data 0.001 (0.073)	Loss 0.486 (0.594)
Epoch: [20][180/200]	Time 0.495 (0.494)	Data 0.000 (0.073)	Loss 0.801 (0.599)
Epoch: [20][200/200]	Time 0.401 (0.493)	Data 0.000 (0.072)	Loss 0.608 (0.602)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.180)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.369718074798584
==> Statistics for epoch 21: 613 clusters
Epoch: [21][20/200]	Time 1.935 (0.533)	Data 1.458 (0.111)	Loss 0.498 (0.108)
Epoch: [21][40/200]	Time 0.553 (0.518)	Data 0.001 (0.092)	Loss 0.746 (0.386)
Epoch: [21][60/200]	Time 0.418 (0.508)	Data 0.001 (0.083)	Loss 0.706 (0.472)
Epoch: [21][80/200]	Time 0.401 (0.502)	Data 0.001 (0.079)	Loss 0.736 (0.518)
Epoch: [21][100/200]	Time 0.409 (0.501)	Data 0.001 (0.077)	Loss 0.530 (0.543)
Epoch: [21][120/200]	Time 0.400 (0.498)	Data 0.001 (0.075)	Loss 0.694 (0.557)
Epoch: [21][140/200]	Time 0.410 (0.496)	Data 0.001 (0.073)	Loss 0.585 (0.562)
Epoch: [21][160/200]	Time 0.414 (0.496)	Data 0.001 (0.072)	Loss 0.471 (0.570)
Epoch: [21][180/200]	Time 0.502 (0.496)	Data 0.001 (0.072)	Loss 0.681 (0.579)
Epoch: [21][200/200]	Time 0.397 (0.495)	Data 0.001 (0.072)	Loss 0.681 (0.587)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.184)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.463218927383423
==> Statistics for epoch 22: 614 clusters
Epoch: [22][20/200]	Time 1.883 (0.530)	Data 1.448 (0.122)	Loss 0.522 (0.104)
Epoch: [22][40/200]	Time 0.419 (0.514)	Data 0.001 (0.096)	Loss 0.783 (0.346)
Epoch: [22][60/200]	Time 0.402 (0.507)	Data 0.001 (0.086)	Loss 0.779 (0.448)
Epoch: [22][80/200]	Time 0.419 (0.500)	Data 0.000 (0.081)	Loss 0.640 (0.498)
Epoch: [22][100/200]	Time 0.402 (0.497)	Data 0.000 (0.077)	Loss 0.743 (0.525)
Epoch: [22][120/200]	Time 0.429 (0.497)	Data 0.001 (0.075)	Loss 1.068 (0.538)
Epoch: [22][140/200]	Time 0.511 (0.495)	Data 0.001 (0.073)	Loss 0.742 (0.551)
Epoch: [22][160/200]	Time 0.398 (0.494)	Data 0.001 (0.073)	Loss 0.759 (0.565)
Epoch: [22][180/200]	Time 0.413 (0.494)	Data 0.001 (0.072)	Loss 0.585 (0.574)
Epoch: [22][200/200]	Time 0.511 (0.493)	Data 0.001 (0.072)	Loss 0.653 (0.581)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.184)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.123284578323364
==> Statistics for epoch 23: 613 clusters
Epoch: [23][20/200]	Time 2.169 (0.541)	Data 1.599 (0.118)	Loss 0.630 (0.117)
Epoch: [23][40/200]	Time 0.421 (0.513)	Data 0.001 (0.093)	Loss 0.664 (0.376)
Epoch: [23][60/200]	Time 0.424 (0.505)	Data 0.001 (0.084)	Loss 0.695 (0.461)
Epoch: [23][80/200]	Time 0.400 (0.502)	Data 0.001 (0.079)	Loss 0.554 (0.491)
Epoch: [23][100/200]	Time 0.540 (0.502)	Data 0.001 (0.078)	Loss 0.886 (0.523)
Epoch: [23][120/200]	Time 0.411 (0.499)	Data 0.001 (0.076)	Loss 0.528 (0.534)
Epoch: [23][140/200]	Time 0.410 (0.498)	Data 0.001 (0.075)	Loss 0.716 (0.543)
Epoch: [23][160/200]	Time 0.398 (0.496)	Data 0.001 (0.074)	Loss 0.481 (0.552)
Epoch: [23][180/200]	Time 0.552 (0.496)	Data 0.001 (0.074)	Loss 0.647 (0.560)
Epoch: [23][200/200]	Time 0.406 (0.495)	Data 0.001 (0.073)	Loss 0.442 (0.567)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.185)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.99583148956299
==> Statistics for epoch 24: 614 clusters
Epoch: [24][20/200]	Time 1.883 (0.531)	Data 1.421 (0.110)	Loss 0.688 (0.114)
Epoch: [24][40/200]	Time 0.414 (0.513)	Data 0.001 (0.093)	Loss 0.743 (0.379)
Epoch: [24][60/200]	Time 0.408 (0.504)	Data 0.001 (0.083)	Loss 0.751 (0.465)
Epoch: [24][80/200]	Time 0.412 (0.501)	Data 0.001 (0.080)	Loss 0.741 (0.500)
Epoch: [24][100/200]	Time 0.410 (0.499)	Data 0.001 (0.079)	Loss 0.818 (0.534)
Epoch: [24][120/200]	Time 0.396 (0.498)	Data 0.001 (0.078)	Loss 0.712 (0.546)
Epoch: [24][140/200]	Time 0.397 (0.496)	Data 0.001 (0.076)	Loss 0.647 (0.559)
Epoch: [24][160/200]	Time 0.523 (0.495)	Data 0.001 (0.075)	Loss 0.762 (0.575)
Epoch: [24][180/200]	Time 0.405 (0.495)	Data 0.001 (0.074)	Loss 0.617 (0.581)
Epoch: [24][200/200]	Time 0.399 (0.494)	Data 0.001 (0.073)	Loss 0.465 (0.585)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.183)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.16594123840332
==> Statistics for epoch 25: 614 clusters
Epoch: [25][20/200]	Time 1.749 (0.539)	Data 1.279 (0.114)	Loss 0.612 (0.109)
Epoch: [25][40/200]	Time 0.401 (0.514)	Data 0.001 (0.093)	Loss 0.592 (0.350)
Epoch: [25][60/200]	Time 0.411 (0.506)	Data 0.001 (0.083)	Loss 0.775 (0.440)
Epoch: [25][80/200]	Time 0.425 (0.502)	Data 0.001 (0.079)	Loss 0.611 (0.486)
Epoch: [25][100/200]	Time 0.414 (0.502)	Data 0.001 (0.079)	Loss 0.689 (0.511)
Epoch: [25][120/200]	Time 0.408 (0.499)	Data 0.001 (0.077)	Loss 0.485 (0.532)
Epoch: [25][140/200]	Time 0.411 (0.498)	Data 0.002 (0.075)	Loss 0.516 (0.547)
Epoch: [25][160/200]	Time 0.408 (0.497)	Data 0.001 (0.074)	Loss 0.420 (0.561)
Epoch: [25][180/200]	Time 0.502 (0.498)	Data 0.002 (0.074)	Loss 0.539 (0.573)
Epoch: [25][200/200]	Time 0.416 (0.496)	Data 0.001 (0.073)	Loss 0.522 (0.577)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.145 (0.187)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.3250253200531
==> Statistics for epoch 26: 613 clusters
Epoch: [26][20/200]	Time 1.978 (0.542)	Data 1.515 (0.111)	Loss 0.641 (0.101)
Epoch: [26][40/200]	Time 0.396 (0.520)	Data 0.001 (0.088)	Loss 0.633 (0.333)
Epoch: [26][60/200]	Time 0.582 (0.508)	Data 0.001 (0.082)	Loss 0.548 (0.426)
Epoch: [26][80/200]	Time 0.412 (0.506)	Data 0.001 (0.081)	Loss 0.612 (0.486)
Epoch: [26][100/200]	Time 0.413 (0.502)	Data 0.001 (0.077)	Loss 0.734 (0.515)
Epoch: [26][120/200]	Time 0.407 (0.500)	Data 0.001 (0.076)	Loss 0.611 (0.536)
Epoch: [26][140/200]	Time 0.412 (0.498)	Data 0.001 (0.075)	Loss 0.465 (0.545)
Epoch: [26][160/200]	Time 0.400 (0.497)	Data 0.001 (0.074)	Loss 0.645 (0.558)
Epoch: [26][180/200]	Time 0.410 (0.496)	Data 0.001 (0.073)	Loss 0.605 (0.565)
Epoch: [26][200/200]	Time 0.407 (0.495)	Data 0.001 (0.072)	Loss 0.576 (0.574)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.145 (0.182)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.986002445220947
==> Statistics for epoch 27: 612 clusters
Epoch: [27][20/200]	Time 2.042 (0.554)	Data 1.542 (0.122)	Loss 0.559 (0.105)
Epoch: [27][40/200]	Time 0.558 (0.526)	Data 0.001 (0.097)	Loss 0.771 (0.346)
Epoch: [27][60/200]	Time 0.400 (0.519)	Data 0.001 (0.093)	Loss 0.691 (0.456)
Epoch: [27][80/200]	Time 0.401 (0.515)	Data 0.001 (0.089)	Loss 0.407 (0.501)
Epoch: [27][100/200]	Time 0.413 (0.511)	Data 0.001 (0.087)	Loss 0.524 (0.530)
Epoch: [27][120/200]	Time 0.401 (0.508)	Data 0.001 (0.085)	Loss 0.511 (0.546)
Epoch: [27][140/200]	Time 0.399 (0.507)	Data 0.001 (0.084)	Loss 0.474 (0.555)
Epoch: [27][160/200]	Time 0.524 (0.507)	Data 0.001 (0.084)	Loss 0.516 (0.559)
Epoch: [27][180/200]	Time 0.414 (0.505)	Data 0.001 (0.083)	Loss 0.871 (0.569)
Epoch: [27][200/200]	Time 0.406 (0.505)	Data 0.001 (0.082)	Loss 0.726 (0.575)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.249 (0.187)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.870758533477783
==> Statistics for epoch 28: 612 clusters
Epoch: [28][20/200]	Time 2.198 (0.551)	Data 1.742 (0.130)	Loss 0.583 (0.106)
Epoch: [28][40/200]	Time 0.431 (0.525)	Data 0.001 (0.100)	Loss 0.502 (0.355)
Epoch: [28][60/200]	Time 0.421 (0.518)	Data 0.000 (0.092)	Loss 0.681 (0.439)
Epoch: [28][80/200]	Time 0.502 (0.511)	Data 0.001 (0.086)	Loss 0.500 (0.480)
Epoch: [28][100/200]	Time 0.408 (0.505)	Data 0.001 (0.082)	Loss 0.762 (0.513)
Epoch: [28][120/200]	Time 0.413 (0.504)	Data 0.001 (0.081)	Loss 0.619 (0.534)
Epoch: [28][140/200]	Time 0.416 (0.503)	Data 0.001 (0.080)	Loss 0.561 (0.545)
Epoch: [28][160/200]	Time 0.398 (0.502)	Data 0.001 (0.079)	Loss 0.652 (0.553)
Epoch: [28][180/200]	Time 0.392 (0.500)	Data 0.001 (0.078)	Loss 0.592 (0.557)
Epoch: [28][200/200]	Time 0.516 (0.501)	Data 0.001 (0.078)	Loss 0.413 (0.562)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.182)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.296037435531616
==> Statistics for epoch 29: 614 clusters
Epoch: [29][20/200]	Time 1.875 (0.516)	Data 1.421 (0.109)	Loss 0.592 (0.102)
Epoch: [29][40/200]	Time 0.422 (0.504)	Data 0.001 (0.086)	Loss 0.781 (0.356)
Epoch: [29][60/200]	Time 0.404 (0.498)	Data 0.001 (0.079)	Loss 0.458 (0.434)
Epoch: [29][80/200]	Time 0.396 (0.491)	Data 0.001 (0.075)	Loss 0.757 (0.480)
Epoch: [29][100/200]	Time 0.413 (0.491)	Data 0.001 (0.074)	Loss 0.712 (0.504)
Epoch: [29][120/200]	Time 0.416 (0.492)	Data 0.001 (0.075)	Loss 0.725 (0.529)
Epoch: [29][140/200]	Time 0.418 (0.492)	Data 0.001 (0.074)	Loss 0.573 (0.544)
Epoch: [29][160/200]	Time 0.410 (0.493)	Data 0.001 (0.074)	Loss 0.786 (0.552)
Epoch: [29][180/200]	Time 0.408 (0.492)	Data 0.001 (0.073)	Loss 0.443 (0.557)
Epoch: [29][200/200]	Time 0.408 (0.491)	Data 0.001 (0.073)	Loss 0.631 (0.566)
Extract Features: [50/76]	Time 0.275 (0.182)	Data 0.000 (0.021)	
Mean AP: 93.6%

 * Finished epoch  29  model mAP: 93.6%  best: 93.6% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.178)	Data 0.000 (0.020)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.114513158798218
==> Statistics for epoch 30: 612 clusters
Epoch: [30][20/200]	Time 1.744 (0.533)	Data 1.280 (0.103)	Loss 0.724 (0.112)
Epoch: [30][40/200]	Time 0.403 (0.509)	Data 0.001 (0.084)	Loss 0.364 (0.332)
Epoch: [30][60/200]	Time 0.402 (0.503)	Data 0.001 (0.083)	Loss 0.424 (0.430)
Epoch: [30][80/200]	Time 0.408 (0.501)	Data 0.001 (0.080)	Loss 0.489 (0.477)
Epoch: [30][100/200]	Time 0.409 (0.498)	Data 0.001 (0.078)	Loss 0.334 (0.510)
Epoch: [30][120/200]	Time 0.528 (0.497)	Data 0.004 (0.076)	Loss 0.650 (0.526)
Epoch: [30][140/200]	Time 0.403 (0.495)	Data 0.001 (0.075)	Loss 0.618 (0.539)
Epoch: [30][160/200]	Time 0.415 (0.498)	Data 0.001 (0.076)	Loss 0.634 (0.548)
Epoch: [30][180/200]	Time 0.507 (0.496)	Data 0.001 (0.075)	Loss 0.657 (0.557)
Epoch: [30][200/200]	Time 0.411 (0.496)	Data 0.001 (0.075)	Loss 0.449 (0.562)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.238 (0.186)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.034758806228638
==> Statistics for epoch 31: 614 clusters
Epoch: [31][20/200]	Time 1.779 (0.531)	Data 1.329 (0.108)	Loss 0.423 (0.092)
Epoch: [31][40/200]	Time 0.441 (0.512)	Data 0.001 (0.087)	Loss 0.586 (0.337)
Epoch: [31][60/200]	Time 0.401 (0.505)	Data 0.000 (0.081)	Loss 0.466 (0.426)
Epoch: [31][80/200]	Time 0.396 (0.500)	Data 0.000 (0.077)	Loss 0.711 (0.483)
Epoch: [31][100/200]	Time 0.399 (0.500)	Data 0.000 (0.077)	Loss 0.560 (0.515)
Epoch: [31][120/200]	Time 0.411 (0.498)	Data 0.000 (0.075)	Loss 0.651 (0.543)
Epoch: [31][140/200]	Time 0.521 (0.496)	Data 0.000 (0.074)	Loss 0.465 (0.554)
Epoch: [31][160/200]	Time 0.408 (0.495)	Data 0.000 (0.073)	Loss 0.519 (0.560)
Epoch: [31][180/200]	Time 0.402 (0.495)	Data 0.000 (0.073)	Loss 0.664 (0.566)
Epoch: [31][200/200]	Time 0.508 (0.494)	Data 0.000 (0.072)	Loss 0.488 (0.566)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.183)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.992727994918823
==> Statistics for epoch 32: 613 clusters
Epoch: [32][20/200]	Time 1.848 (0.540)	Data 1.406 (0.110)	Loss 0.661 (0.116)
Epoch: [32][40/200]	Time 0.401 (0.516)	Data 0.001 (0.091)	Loss 0.516 (0.362)
Epoch: [32][60/200]	Time 0.411 (0.506)	Data 0.001 (0.084)	Loss 0.769 (0.448)
Epoch: [32][80/200]	Time 0.397 (0.502)	Data 0.001 (0.079)	Loss 0.666 (0.495)
Epoch: [32][100/200]	Time 0.413 (0.500)	Data 0.000 (0.077)	Loss 0.475 (0.522)
Epoch: [32][120/200]	Time 0.424 (0.499)	Data 0.000 (0.076)	Loss 0.512 (0.531)
Epoch: [32][140/200]	Time 0.402 (0.497)	Data 0.000 (0.075)	Loss 0.704 (0.546)
Epoch: [32][160/200]	Time 0.409 (0.495)	Data 0.000 (0.073)	Loss 0.594 (0.551)
Epoch: [32][180/200]	Time 0.400 (0.494)	Data 0.000 (0.072)	Loss 0.591 (0.560)
Epoch: [32][200/200]	Time 0.410 (0.493)	Data 0.000 (0.071)	Loss 0.389 (0.570)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.145 (0.182)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.215426683425903
==> Statistics for epoch 33: 612 clusters
Epoch: [33][20/200]	Time 1.731 (0.523)	Data 1.266 (0.100)	Loss 0.427 (0.099)
Epoch: [33][40/200]	Time 0.396 (0.507)	Data 0.001 (0.084)	Loss 0.546 (0.343)
Epoch: [33][60/200]	Time 0.516 (0.503)	Data 0.001 (0.079)	Loss 0.610 (0.427)
Epoch: [33][80/200]	Time 0.399 (0.502)	Data 0.001 (0.080)	Loss 0.608 (0.470)
Epoch: [33][100/200]	Time 0.420 (0.502)	Data 0.001 (0.078)	Loss 1.016 (0.499)
Epoch: [33][120/200]	Time 0.514 (0.500)	Data 0.001 (0.076)	Loss 0.887 (0.512)
Epoch: [33][140/200]	Time 0.411 (0.497)	Data 0.001 (0.075)	Loss 0.537 (0.529)
Epoch: [33][160/200]	Time 0.413 (0.496)	Data 0.001 (0.073)	Loss 0.535 (0.544)
Epoch: [33][180/200]	Time 0.410 (0.496)	Data 0.001 (0.072)	Loss 0.488 (0.554)
Epoch: [33][200/200]	Time 0.398 (0.494)	Data 0.002 (0.072)	Loss 0.835 (0.561)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.271 (0.191)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.180454969406128
==> Statistics for epoch 34: 612 clusters
Epoch: [34][20/200]	Time 2.053 (0.542)	Data 1.603 (0.118)	Loss 0.731 (0.120)
Epoch: [34][40/200]	Time 0.429 (0.515)	Data 0.001 (0.091)	Loss 0.709 (0.354)
Epoch: [34][60/200]	Time 0.419 (0.513)	Data 0.001 (0.087)	Loss 0.574 (0.430)
Epoch: [34][80/200]	Time 0.514 (0.509)	Data 0.001 (0.085)	Loss 0.572 (0.473)
Epoch: [34][100/200]	Time 0.410 (0.505)	Data 0.001 (0.082)	Loss 0.773 (0.509)
Epoch: [34][120/200]	Time 0.405 (0.504)	Data 0.001 (0.080)	Loss 0.947 (0.524)
Epoch: [34][140/200]	Time 0.411 (0.503)	Data 0.001 (0.080)	Loss 0.437 (0.538)
Epoch: [34][160/200]	Time 0.421 (0.503)	Data 0.001 (0.079)	Loss 0.687 (0.549)
Epoch: [34][180/200]	Time 0.439 (0.503)	Data 0.001 (0.079)	Loss 0.777 (0.554)
Epoch: [34][200/200]	Time 0.505 (0.503)	Data 0.001 (0.079)	Loss 0.468 (0.556)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.186)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.026880502700806
==> Statistics for epoch 35: 612 clusters
Epoch: [35][20/200]	Time 1.704 (0.526)	Data 1.244 (0.104)	Loss 0.507 (0.099)
Epoch: [35][40/200]	Time 0.415 (0.506)	Data 0.001 (0.087)	Loss 0.607 (0.348)
Epoch: [35][60/200]	Time 0.420 (0.502)	Data 0.001 (0.081)	Loss 0.618 (0.449)
Epoch: [35][80/200]	Time 0.431 (0.501)	Data 0.001 (0.079)	Loss 0.484 (0.480)
Epoch: [35][100/200]	Time 0.517 (0.498)	Data 0.001 (0.075)	Loss 0.868 (0.508)
Epoch: [35][120/200]	Time 0.409 (0.497)	Data 0.001 (0.075)	Loss 0.497 (0.522)
Epoch: [35][140/200]	Time 0.402 (0.496)	Data 0.001 (0.074)	Loss 0.545 (0.533)
Epoch: [35][160/200]	Time 0.407 (0.494)	Data 0.001 (0.073)	Loss 0.591 (0.548)
Epoch: [35][180/200]	Time 0.396 (0.494)	Data 0.001 (0.073)	Loss 0.747 (0.554)
Epoch: [35][200/200]	Time 0.423 (0.496)	Data 0.001 (0.074)	Loss 0.532 (0.556)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.150 (0.181)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.08660078048706
==> Statistics for epoch 36: 612 clusters
Epoch: [36][20/200]	Time 1.900 (0.533)	Data 1.415 (0.108)	Loss 0.365 (0.095)
Epoch: [36][40/200]	Time 0.411 (0.514)	Data 0.001 (0.090)	Loss 0.512 (0.340)
Epoch: [36][60/200]	Time 0.528 (0.506)	Data 0.001 (0.084)	Loss 0.517 (0.424)
Epoch: [36][80/200]	Time 0.410 (0.502)	Data 0.001 (0.080)	Loss 0.643 (0.479)
Epoch: [36][100/200]	Time 0.404 (0.500)	Data 0.001 (0.079)	Loss 0.645 (0.505)
Epoch: [36][120/200]	Time 0.408 (0.499)	Data 0.001 (0.077)	Loss 0.802 (0.520)
Epoch: [36][140/200]	Time 0.416 (0.497)	Data 0.001 (0.075)	Loss 0.614 (0.530)
Epoch: [36][160/200]	Time 0.397 (0.496)	Data 0.001 (0.074)	Loss 0.510 (0.543)
Epoch: [36][180/200]	Time 0.402 (0.495)	Data 0.001 (0.074)	Loss 0.870 (0.553)
Epoch: [36][200/200]	Time 0.419 (0.495)	Data 0.001 (0.073)	Loss 0.607 (0.556)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.182)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.235575437545776
==> Statistics for epoch 37: 613 clusters
Epoch: [37][20/200]	Time 1.904 (0.540)	Data 1.419 (0.113)	Loss 0.583 (0.103)
Epoch: [37][40/200]	Time 0.400 (0.522)	Data 0.001 (0.096)	Loss 0.690 (0.355)
Epoch: [37][60/200]	Time 0.407 (0.513)	Data 0.001 (0.090)	Loss 0.722 (0.448)
Epoch: [37][80/200]	Time 0.397 (0.509)	Data 0.001 (0.085)	Loss 0.391 (0.494)
Epoch: [37][100/200]	Time 0.399 (0.506)	Data 0.001 (0.083)	Loss 0.675 (0.512)
Epoch: [37][120/200]	Time 0.528 (0.507)	Data 0.001 (0.083)	Loss 0.798 (0.528)
Epoch: [37][140/200]	Time 0.395 (0.504)	Data 0.002 (0.081)	Loss 0.440 (0.533)
Epoch: [37][160/200]	Time 0.417 (0.502)	Data 0.001 (0.079)	Loss 0.454 (0.544)
Epoch: [37][180/200]	Time 0.394 (0.502)	Data 0.001 (0.079)	Loss 0.565 (0.552)
Epoch: [37][200/200]	Time 0.398 (0.502)	Data 0.001 (0.079)	Loss 0.804 (0.556)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.182)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.40327501296997
==> Statistics for epoch 38: 612 clusters
Epoch: [38][20/200]	Time 1.998 (0.540)	Data 1.440 (0.116)	Loss 0.505 (0.098)
Epoch: [38][40/200]	Time 0.416 (0.511)	Data 0.001 (0.095)	Loss 0.673 (0.340)
Epoch: [38][60/200]	Time 0.416 (0.505)	Data 0.001 (0.088)	Loss 0.675 (0.425)
Epoch: [38][80/200]	Time 0.408 (0.502)	Data 0.001 (0.082)	Loss 0.721 (0.478)
Epoch: [38][100/200]	Time 0.511 (0.499)	Data 0.001 (0.081)	Loss 0.840 (0.501)
Epoch: [38][120/200]	Time 0.415 (0.499)	Data 0.001 (0.081)	Loss 0.518 (0.526)
Epoch: [38][140/200]	Time 0.416 (0.499)	Data 0.001 (0.079)	Loss 0.668 (0.538)
Epoch: [38][160/200]	Time 0.407 (0.499)	Data 0.001 (0.079)	Loss 0.564 (0.548)
Epoch: [38][180/200]	Time 0.410 (0.499)	Data 0.001 (0.079)	Loss 0.615 (0.555)
Epoch: [38][200/200]	Time 0.408 (0.499)	Data 0.001 (0.079)	Loss 0.414 (0.556)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.149 (0.182)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.77338933944702
==> Statistics for epoch 39: 611 clusters
Epoch: [39][20/200]	Time 1.999 (0.554)	Data 1.528 (0.125)	Loss 0.402 (0.092)
Epoch: [39][40/200]	Time 0.398 (0.523)	Data 0.001 (0.097)	Loss 0.522 (0.335)
Epoch: [39][60/200]	Time 0.536 (0.517)	Data 0.001 (0.091)	Loss 0.321 (0.434)
Epoch: [39][80/200]	Time 0.412 (0.509)	Data 0.001 (0.085)	Loss 0.742 (0.475)
Epoch: [39][100/200]	Time 0.402 (0.508)	Data 0.001 (0.083)	Loss 0.440 (0.495)
Epoch: [39][120/200]	Time 0.403 (0.506)	Data 0.001 (0.080)	Loss 0.639 (0.505)
Epoch: [39][140/200]	Time 0.395 (0.504)	Data 0.001 (0.080)	Loss 0.465 (0.519)
Epoch: [39][160/200]	Time 0.419 (0.504)	Data 0.001 (0.080)	Loss 0.474 (0.529)
Epoch: [39][180/200]	Time 0.412 (0.502)	Data 0.001 (0.078)	Loss 0.553 (0.535)
Epoch: [39][200/200]	Time 0.416 (0.501)	Data 0.001 (0.078)	Loss 0.528 (0.541)
Extract Features: [50/76]	Time 0.143 (0.185)	Data 0.000 (0.025)	
Mean AP: 93.7%

 * Finished epoch  39  model mAP: 93.7%  best: 93.7% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.142 (0.181)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.238273859024048
==> Statistics for epoch 40: 612 clusters
Epoch: [40][20/200]	Time 1.791 (0.528)	Data 1.342 (0.104)	Loss 0.641 (0.102)
Epoch: [40][40/200]	Time 0.422 (0.510)	Data 0.001 (0.083)	Loss 0.580 (0.340)
Epoch: [40][60/200]	Time 0.564 (0.503)	Data 0.001 (0.077)	Loss 0.494 (0.415)
Epoch: [40][80/200]	Time 0.414 (0.500)	Data 0.001 (0.075)	Loss 0.541 (0.460)
Epoch: [40][100/200]	Time 0.404 (0.498)	Data 0.001 (0.074)	Loss 0.467 (0.499)
Epoch: [40][120/200]	Time 0.414 (0.495)	Data 0.001 (0.072)	Loss 0.537 (0.525)
Epoch: [40][140/200]	Time 0.412 (0.496)	Data 0.001 (0.073)	Loss 0.613 (0.533)
Epoch: [40][160/200]	Time 0.406 (0.496)	Data 0.001 (0.073)	Loss 0.455 (0.535)
Epoch: [40][180/200]	Time 0.406 (0.496)	Data 0.001 (0.072)	Loss 0.593 (0.542)
Epoch: [40][200/200]	Time 0.415 (0.495)	Data 0.003 (0.071)	Loss 0.712 (0.546)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.182)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.772342205047607
==> Statistics for epoch 41: 611 clusters
Epoch: [41][20/200]	Time 1.979 (0.541)	Data 1.475 (0.116)	Loss 0.680 (0.112)
Epoch: [41][40/200]	Time 0.425 (0.522)	Data 0.001 (0.097)	Loss 0.620 (0.348)
Epoch: [41][60/200]	Time 0.408 (0.514)	Data 0.001 (0.087)	Loss 0.455 (0.427)
Epoch: [41][80/200]	Time 0.538 (0.512)	Data 0.001 (0.085)	Loss 0.550 (0.470)
Epoch: [41][100/200]	Time 0.397 (0.508)	Data 0.001 (0.084)	Loss 0.422 (0.499)
Epoch: [41][120/200]	Time 0.399 (0.505)	Data 0.002 (0.082)	Loss 0.567 (0.518)
Epoch: [41][140/200]	Time 0.411 (0.504)	Data 0.001 (0.081)	Loss 0.546 (0.531)
Epoch: [41][160/200]	Time 0.409 (0.504)	Data 0.001 (0.080)	Loss 0.603 (0.541)
Epoch: [41][180/200]	Time 0.491 (0.503)	Data 0.001 (0.080)	Loss 0.488 (0.550)
Epoch: [41][200/200]	Time 0.414 (0.502)	Data 0.001 (0.079)	Loss 0.570 (0.557)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.188)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.237502574920654
==> Statistics for epoch 42: 612 clusters
Epoch: [42][20/200]	Time 1.748 (0.531)	Data 1.291 (0.102)	Loss 0.378 (0.095)
Epoch: [42][40/200]	Time 0.406 (0.506)	Data 0.001 (0.084)	Loss 0.675 (0.343)
Epoch: [42][60/200]	Time 0.399 (0.502)	Data 0.001 (0.079)	Loss 0.593 (0.439)
Epoch: [42][80/200]	Time 0.405 (0.498)	Data 0.001 (0.075)	Loss 0.622 (0.479)
Epoch: [42][100/200]	Time 0.414 (0.495)	Data 0.001 (0.074)	Loss 0.520 (0.511)
Epoch: [42][120/200]	Time 0.400 (0.494)	Data 0.001 (0.072)	Loss 0.479 (0.528)
Epoch: [42][140/200]	Time 0.416 (0.494)	Data 0.001 (0.072)	Loss 0.594 (0.540)
Epoch: [42][160/200]	Time 0.415 (0.493)	Data 0.001 (0.071)	Loss 0.668 (0.549)
Epoch: [42][180/200]	Time 0.410 (0.493)	Data 0.001 (0.071)	Loss 0.392 (0.561)
Epoch: [42][200/200]	Time 0.410 (0.493)	Data 0.001 (0.071)	Loss 0.778 (0.571)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.179)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.418816328048706
==> Statistics for epoch 43: 614 clusters
Epoch: [43][20/200]	Time 1.925 (0.543)	Data 1.483 (0.114)	Loss 0.535 (0.097)
Epoch: [43][40/200]	Time 0.542 (0.523)	Data 0.001 (0.095)	Loss 0.556 (0.337)
Epoch: [43][60/200]	Time 0.415 (0.513)	Data 0.001 (0.087)	Loss 0.788 (0.418)
Epoch: [43][80/200]	Time 0.398 (0.512)	Data 0.001 (0.084)	Loss 0.530 (0.459)
Epoch: [43][100/200]	Time 0.412 (0.509)	Data 0.001 (0.081)	Loss 0.589 (0.482)
Epoch: [43][120/200]	Time 0.403 (0.506)	Data 0.001 (0.080)	Loss 0.565 (0.506)
Epoch: [43][140/200]	Time 0.417 (0.506)	Data 0.001 (0.079)	Loss 0.669 (0.523)
Epoch: [43][160/200]	Time 0.507 (0.505)	Data 0.001 (0.078)	Loss 0.836 (0.531)
Epoch: [43][180/200]	Time 0.413 (0.504)	Data 0.001 (0.078)	Loss 0.595 (0.535)
Epoch: [43][200/200]	Time 0.411 (0.502)	Data 0.001 (0.077)	Loss 0.554 (0.541)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.189)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.11159038543701
==> Statistics for epoch 44: 612 clusters
Epoch: [44][20/200]	Time 1.892 (0.544)	Data 1.337 (0.119)	Loss 0.597 (0.108)
Epoch: [44][40/200]	Time 0.418 (0.512)	Data 0.003 (0.092)	Loss 1.051 (0.351)
Epoch: [44][60/200]	Time 0.406 (0.504)	Data 0.001 (0.085)	Loss 0.712 (0.433)
Epoch: [44][80/200]	Time 0.408 (0.501)	Data 0.001 (0.080)	Loss 0.431 (0.464)
Epoch: [44][100/200]	Time 0.413 (0.497)	Data 0.001 (0.077)	Loss 0.778 (0.496)
Epoch: [44][120/200]	Time 0.413 (0.496)	Data 0.001 (0.076)	Loss 0.471 (0.507)
Epoch: [44][140/200]	Time 0.391 (0.494)	Data 0.001 (0.074)	Loss 0.581 (0.516)
Epoch: [44][160/200]	Time 0.393 (0.492)	Data 0.001 (0.073)	Loss 0.577 (0.522)
Epoch: [44][180/200]	Time 0.393 (0.492)	Data 0.001 (0.073)	Loss 0.457 (0.530)
Epoch: [44][200/200]	Time 0.397 (0.491)	Data 0.001 (0.073)	Loss 0.783 (0.536)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.180)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.77820587158203
==> Statistics for epoch 45: 614 clusters
Epoch: [45][20/200]	Time 1.842 (0.532)	Data 1.387 (0.106)	Loss 0.787 (0.113)
Epoch: [45][40/200]	Time 0.420 (0.512)	Data 0.001 (0.086)	Loss 0.588 (0.342)
Epoch: [45][60/200]	Time 0.407 (0.504)	Data 0.001 (0.080)	Loss 0.548 (0.427)
Epoch: [45][80/200]	Time 0.406 (0.499)	Data 0.001 (0.075)	Loss 0.565 (0.479)
Epoch: [45][100/200]	Time 0.396 (0.498)	Data 0.001 (0.075)	Loss 0.740 (0.504)
Epoch: [45][120/200]	Time 0.512 (0.496)	Data 0.001 (0.073)	Loss 0.645 (0.518)
Epoch: [45][140/200]	Time 0.404 (0.495)	Data 0.001 (0.073)	Loss 0.517 (0.523)
Epoch: [45][160/200]	Time 0.402 (0.495)	Data 0.001 (0.072)	Loss 0.425 (0.529)
Epoch: [45][180/200]	Time 0.410 (0.494)	Data 0.001 (0.072)	Loss 0.598 (0.536)
Epoch: [45][200/200]	Time 0.411 (0.493)	Data 0.001 (0.071)	Loss 0.679 (0.547)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.147 (0.185)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.572314023971558
==> Statistics for epoch 46: 612 clusters
Epoch: [46][20/200]	Time 2.044 (0.539)	Data 1.601 (0.113)	Loss 0.830 (0.112)
Epoch: [46][40/200]	Time 0.509 (0.514)	Data 0.001 (0.090)	Loss 0.512 (0.327)
Epoch: [46][60/200]	Time 0.411 (0.504)	Data 0.001 (0.084)	Loss 0.306 (0.403)
Epoch: [46][80/200]	Time 0.412 (0.501)	Data 0.001 (0.081)	Loss 0.327 (0.456)
Epoch: [46][100/200]	Time 0.410 (0.498)	Data 0.001 (0.078)	Loss 0.635 (0.484)
Epoch: [46][120/200]	Time 0.417 (0.497)	Data 0.001 (0.077)	Loss 0.785 (0.504)
Epoch: [46][140/200]	Time 0.400 (0.496)	Data 0.001 (0.076)	Loss 0.490 (0.521)
Epoch: [46][160/200]	Time 0.403 (0.495)	Data 0.001 (0.075)	Loss 0.518 (0.534)
Epoch: [46][180/200]	Time 0.395 (0.494)	Data 0.001 (0.074)	Loss 0.397 (0.537)
Epoch: [46][200/200]	Time 0.409 (0.493)	Data 0.001 (0.073)	Loss 0.492 (0.541)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.231 (0.185)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 18.99992609024048
==> Statistics for epoch 47: 611 clusters
Epoch: [47][20/200]	Time 1.762 (0.534)	Data 1.282 (0.104)	Loss 0.667 (0.112)
Epoch: [47][40/200]	Time 0.417 (0.512)	Data 0.001 (0.084)	Loss 0.788 (0.348)
Epoch: [47][60/200]	Time 0.420 (0.507)	Data 0.001 (0.078)	Loss 0.680 (0.420)
Epoch: [47][80/200]	Time 0.530 (0.506)	Data 0.001 (0.076)	Loss 0.635 (0.464)
Epoch: [47][100/200]	Time 0.406 (0.502)	Data 0.001 (0.074)	Loss 0.633 (0.491)
Epoch: [47][120/200]	Time 0.405 (0.500)	Data 0.001 (0.072)	Loss 0.592 (0.503)
Epoch: [47][140/200]	Time 0.411 (0.497)	Data 0.001 (0.071)	Loss 0.416 (0.512)
Epoch: [47][160/200]	Time 0.409 (0.497)	Data 0.002 (0.071)	Loss 0.501 (0.523)
Epoch: [47][180/200]	Time 0.397 (0.498)	Data 0.001 (0.073)	Loss 0.688 (0.537)
Epoch: [47][200/200]	Time 0.403 (0.498)	Data 0.001 (0.073)	Loss 0.605 (0.541)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.184)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.401251316070557
==> Statistics for epoch 48: 614 clusters
Epoch: [48][20/200]	Time 2.074 (0.551)	Data 1.614 (0.118)	Loss 0.705 (0.107)
Epoch: [48][40/200]	Time 0.423 (0.527)	Data 0.001 (0.099)	Loss 0.493 (0.344)
Epoch: [48][60/200]	Time 0.417 (0.518)	Data 0.001 (0.093)	Loss 0.369 (0.425)
Epoch: [48][80/200]	Time 0.405 (0.513)	Data 0.001 (0.088)	Loss 0.632 (0.454)
Epoch: [48][100/200]	Time 0.401 (0.509)	Data 0.002 (0.086)	Loss 0.545 (0.481)
Epoch: [48][120/200]	Time 0.555 (0.507)	Data 0.001 (0.083)	Loss 0.582 (0.505)
Epoch: [48][140/200]	Time 0.398 (0.504)	Data 0.001 (0.082)	Loss 0.529 (0.518)
Epoch: [48][160/200]	Time 0.395 (0.503)	Data 0.001 (0.082)	Loss 0.670 (0.527)
Epoch: [48][180/200]	Time 0.413 (0.504)	Data 0.001 (0.082)	Loss 0.812 (0.535)
Epoch: [48][200/200]	Time 0.412 (0.503)	Data 0.001 (0.081)	Loss 0.645 (0.539)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.145 (0.187)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.369293689727783
==> Statistics for epoch 49: 612 clusters
Epoch: [49][20/200]	Time 1.882 (0.542)	Data 1.404 (0.115)	Loss 0.526 (0.096)
Epoch: [49][40/200]	Time 0.427 (0.522)	Data 0.001 (0.096)	Loss 0.335 (0.323)
Epoch: [49][60/200]	Time 0.403 (0.514)	Data 0.001 (0.088)	Loss 0.484 (0.412)
Epoch: [49][80/200]	Time 0.398 (0.511)	Data 0.001 (0.085)	Loss 0.546 (0.453)
Epoch: [49][100/200]	Time 0.411 (0.509)	Data 0.001 (0.084)	Loss 0.571 (0.487)
Epoch: [49][120/200]	Time 0.414 (0.506)	Data 0.001 (0.082)	Loss 0.400 (0.502)
Epoch: [49][140/200]	Time 0.394 (0.504)	Data 0.001 (0.080)	Loss 0.665 (0.522)
Epoch: [49][160/200]	Time 0.402 (0.501)	Data 0.001 (0.078)	Loss 0.525 (0.528)
Epoch: [49][180/200]	Time 0.404 (0.499)	Data 0.002 (0.077)	Loss 0.543 (0.533)
Epoch: [49][200/200]	Time 0.395 (0.499)	Data 0.001 (0.077)	Loss 0.621 (0.536)
Extract Features: [50/76]	Time 0.152 (0.188)	Data 0.000 (0.030)	
Mean AP: 93.6%

 * Finished epoch  49  model mAP: 93.6%  best: 93.7%

==> Test with the best model:
=> Loaded checkpoint 'log/msmt2market/resnet101_ibn_cion/model_best.pth.tar'
Extract Features: [50/76]	Time 0.143 (0.179)	Data 0.000 (0.021)	
Mean AP: 93.7%
CMC Scores:
  top-1          96.8%
  top-5          99.0%
  top-10         99.4%
Total running time:  1:52:02.199995
