==========
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_ibn50a', pretrained_path='/home/ma-user/work/Projects/ReIDNet_Finetune/TransReID/log/bs_exp/ResNet50_IBN_MSMT17/64bs_lr0.0004_ep120_warm20_seed0/resnet50_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/resnet50_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 0.092 (0.476)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 21.796549081802368
Clustering criterion: eps: 0.600
==> Statistics for epoch 0: 540 clusters
Epoch: [0][20/200]	Time 0.257 (0.667)	Data 0.001 (0.099)	Loss 3.326 (3.276)
Epoch: [0][40/200]	Time 0.258 (0.499)	Data 0.001 (0.082)	Loss 2.562 (3.215)
Epoch: [0][60/200]	Time 0.256 (0.443)	Data 0.000 (0.077)	Loss 2.246 (2.953)
Epoch: [0][80/200]	Time 0.256 (0.414)	Data 0.000 (0.073)	Loss 1.889 (2.780)
Epoch: [0][100/200]	Time 0.263 (0.409)	Data 0.001 (0.083)	Loss 1.561 (2.642)
Epoch: [0][120/200]	Time 0.259 (0.394)	Data 0.001 (0.080)	Loss 1.793 (2.543)
Epoch: [0][140/200]	Time 0.257 (0.385)	Data 0.000 (0.077)	Loss 1.699 (2.451)
Epoch: [0][160/200]	Time 0.257 (0.378)	Data 0.000 (0.075)	Loss 2.238 (2.386)
Epoch: [0][180/200]	Time 0.257 (0.379)	Data 0.001 (0.081)	Loss 1.560 (2.321)
Epoch: [0][200/200]	Time 0.257 (0.374)	Data 0.001 (0.079)	Loss 2.188 (2.273)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.146)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.856451272964478
==> Statistics for epoch 1: 580 clusters
Epoch: [1][20/200]	Time 0.265 (0.368)	Data 0.001 (0.099)	Loss 1.579 (0.457)
Epoch: [1][40/200]	Time 0.260 (0.347)	Data 0.001 (0.079)	Loss 2.357 (1.158)
Epoch: [1][60/200]	Time 0.263 (0.342)	Data 0.001 (0.074)	Loss 2.333 (1.396)
Epoch: [1][80/200]	Time 0.258 (0.338)	Data 0.001 (0.072)	Loss 1.626 (1.484)
Epoch: [1][100/200]	Time 0.261 (0.337)	Data 0.001 (0.070)	Loss 1.518 (1.564)
Epoch: [1][120/200]	Time 0.257 (0.335)	Data 0.000 (0.068)	Loss 1.548 (1.590)
Epoch: [1][140/200]	Time 0.258 (0.335)	Data 0.000 (0.068)	Loss 1.939 (1.609)
Epoch: [1][160/200]	Time 0.257 (0.334)	Data 0.000 (0.067)	Loss 1.763 (1.627)
Epoch: [1][180/200]	Time 0.258 (0.334)	Data 0.000 (0.067)	Loss 1.570 (1.642)
Epoch: [1][200/200]	Time 0.258 (0.340)	Data 0.001 (0.073)	Loss 1.400 (1.653)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.092 (0.132)	Data 0.000 (0.034)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.9846248626709
==> Statistics for epoch 2: 599 clusters
Epoch: [2][20/200]	Time 0.256 (0.362)	Data 0.001 (0.102)	Loss 1.337 (0.416)
Epoch: [2][40/200]	Time 0.260 (0.347)	Data 0.001 (0.084)	Loss 1.488 (1.063)
Epoch: [2][60/200]	Time 0.258 (0.342)	Data 0.001 (0.078)	Loss 2.127 (1.267)
Epoch: [2][80/200]	Time 0.254 (0.339)	Data 0.001 (0.076)	Loss 1.864 (1.384)
Epoch: [2][100/200]	Time 0.263 (0.337)	Data 0.001 (0.074)	Loss 1.958 (1.434)
Epoch: [2][120/200]	Time 0.262 (0.337)	Data 0.000 (0.074)	Loss 1.179 (1.480)
Epoch: [2][140/200]	Time 0.255 (0.336)	Data 0.000 (0.072)	Loss 1.809 (1.499)
Epoch: [2][160/200]	Time 0.255 (0.336)	Data 0.000 (0.072)	Loss 1.005 (1.512)
Epoch: [2][180/200]	Time 0.255 (0.335)	Data 0.000 (0.071)	Loss 1.552 (1.526)
Epoch: [2][200/200]	Time 0.267 (0.342)	Data 0.001 (0.078)	Loss 1.774 (1.529)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.135)	Data 0.000 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.594343900680542
==> Statistics for epoch 3: 603 clusters
Epoch: [3][20/200]	Time 0.258 (0.367)	Data 0.001 (0.097)	Loss 1.996 (0.424)
Epoch: [3][40/200]	Time 0.266 (0.350)	Data 0.001 (0.083)	Loss 1.314 (0.945)
Epoch: [3][60/200]	Time 0.258 (0.350)	Data 0.001 (0.083)	Loss 1.706 (1.139)
Epoch: [3][80/200]	Time 0.264 (0.345)	Data 0.001 (0.077)	Loss 1.548 (1.227)
Epoch: [3][100/200]	Time 0.257 (0.344)	Data 0.001 (0.077)	Loss 1.286 (1.296)
Epoch: [3][120/200]	Time 0.260 (0.342)	Data 0.000 (0.075)	Loss 1.614 (1.335)
Epoch: [3][140/200]	Time 0.256 (0.340)	Data 0.000 (0.074)	Loss 1.545 (1.365)
Epoch: [3][160/200]	Time 0.260 (0.338)	Data 0.000 (0.072)	Loss 1.433 (1.382)
Epoch: [3][180/200]	Time 0.261 (0.337)	Data 0.000 (0.072)	Loss 1.116 (1.395)
Epoch: [3][200/200]	Time 0.261 (0.344)	Data 0.001 (0.078)	Loss 1.442 (1.408)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.102 (0.130)	Data 0.006 (0.033)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.649457931518555
==> Statistics for epoch 4: 604 clusters
Epoch: [4][20/200]	Time 0.270 (0.373)	Data 0.001 (0.112)	Loss 1.351 (0.338)
Epoch: [4][40/200]	Time 0.308 (0.355)	Data 0.001 (0.091)	Loss 1.068 (0.854)
Epoch: [4][60/200]	Time 0.256 (0.350)	Data 0.001 (0.087)	Loss 1.405 (1.050)
Epoch: [4][80/200]	Time 0.269 (0.346)	Data 0.001 (0.083)	Loss 1.376 (1.146)
Epoch: [4][100/200]	Time 0.259 (0.344)	Data 0.001 (0.080)	Loss 1.449 (1.204)
Epoch: [4][120/200]	Time 0.259 (0.343)	Data 0.000 (0.079)	Loss 1.264 (1.231)
Epoch: [4][140/200]	Time 0.257 (0.342)	Data 0.000 (0.078)	Loss 1.484 (1.256)
Epoch: [4][160/200]	Time 0.257 (0.342)	Data 0.000 (0.077)	Loss 1.523 (1.277)
Epoch: [4][180/200]	Time 0.257 (0.341)	Data 0.000 (0.077)	Loss 1.280 (1.291)
Epoch: [4][200/200]	Time 0.269 (0.349)	Data 0.001 (0.085)	Loss 1.504 (1.306)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.096 (0.141)	Data 0.000 (0.041)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.291937828063965
==> Statistics for epoch 5: 613 clusters
Epoch: [5][20/200]	Time 1.815 (0.391)	Data 1.547 (0.123)	Loss 1.518 (0.293)
Epoch: [5][40/200]	Time 0.258 (0.361)	Data 0.001 (0.097)	Loss 1.711 (0.847)
Epoch: [5][60/200]	Time 0.257 (0.356)	Data 0.001 (0.091)	Loss 1.383 (1.012)
Epoch: [5][80/200]	Time 0.382 (0.354)	Data 0.001 (0.089)	Loss 1.549 (1.103)
Epoch: [5][100/200]	Time 0.305 (0.349)	Data 0.001 (0.084)	Loss 1.289 (1.164)
Epoch: [5][120/200]	Time 0.257 (0.347)	Data 0.001 (0.083)	Loss 1.236 (1.209)
Epoch: [5][140/200]	Time 0.258 (0.346)	Data 0.001 (0.081)	Loss 1.804 (1.239)
Epoch: [5][160/200]	Time 0.271 (0.345)	Data 0.001 (0.080)	Loss 0.972 (1.248)
Epoch: [5][180/200]	Time 0.260 (0.344)	Data 0.001 (0.079)	Loss 1.146 (1.257)
Epoch: [5][200/200]	Time 0.259 (0.344)	Data 0.001 (0.079)	Loss 1.593 (1.268)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.146 (0.142)	Data 0.053 (0.043)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.55682682991028
==> Statistics for epoch 6: 612 clusters
Epoch: [6][20/200]	Time 1.715 (0.372)	Data 1.404 (0.107)	Loss 1.371 (0.256)
Epoch: [6][40/200]	Time 0.261 (0.348)	Data 0.001 (0.085)	Loss 1.372 (0.761)
Epoch: [6][60/200]	Time 0.261 (0.343)	Data 0.001 (0.079)	Loss 1.327 (0.927)
Epoch: [6][80/200]	Time 0.259 (0.341)	Data 0.000 (0.077)	Loss 1.343 (1.022)
Epoch: [6][100/200]	Time 0.264 (0.339)	Data 0.001 (0.074)	Loss 1.133 (1.070)
Epoch: [6][120/200]	Time 0.256 (0.339)	Data 0.001 (0.074)	Loss 1.368 (1.106)
Epoch: [6][140/200]	Time 0.267 (0.339)	Data 0.001 (0.074)	Loss 1.066 (1.130)
Epoch: [6][160/200]	Time 0.264 (0.340)	Data 0.001 (0.074)	Loss 1.207 (1.151)
Epoch: [6][180/200]	Time 0.263 (0.340)	Data 0.001 (0.074)	Loss 1.363 (1.167)
Epoch: [6][200/200]	Time 0.269 (0.340)	Data 0.001 (0.074)	Loss 1.239 (1.178)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.092 (0.142)	Data 0.000 (0.043)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.626683712005615
==> Statistics for epoch 7: 615 clusters
Epoch: [7][20/200]	Time 1.588 (0.369)	Data 1.298 (0.106)	Loss 1.196 (0.235)
Epoch: [7][40/200]	Time 0.258 (0.348)	Data 0.001 (0.084)	Loss 1.359 (0.720)
Epoch: [7][60/200]	Time 0.390 (0.342)	Data 0.001 (0.076)	Loss 1.470 (0.906)
Epoch: [7][80/200]	Time 0.276 (0.340)	Data 0.001 (0.074)	Loss 1.124 (0.992)
Epoch: [7][100/200]	Time 0.263 (0.338)	Data 0.001 (0.072)	Loss 1.407 (1.029)
Epoch: [7][120/200]	Time 0.265 (0.337)	Data 0.001 (0.071)	Loss 1.462 (1.069)
Epoch: [7][140/200]	Time 0.256 (0.335)	Data 0.001 (0.070)	Loss 0.938 (1.084)
Epoch: [7][160/200]	Time 0.260 (0.335)	Data 0.001 (0.070)	Loss 1.433 (1.111)
Epoch: [7][180/200]	Time 0.258 (0.335)	Data 0.001 (0.070)	Loss 1.260 (1.118)
Epoch: [7][200/200]	Time 0.267 (0.334)	Data 0.001 (0.069)	Loss 1.203 (1.134)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.134)	Data 0.000 (0.036)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.233561038970947
==> Statistics for epoch 8: 614 clusters
Epoch: [8][20/200]	Time 1.575 (0.372)	Data 1.295 (0.107)	Loss 1.148 (0.219)
Epoch: [8][40/200]	Time 0.255 (0.349)	Data 0.001 (0.084)	Loss 1.400 (0.678)
Epoch: [8][60/200]	Time 0.259 (0.342)	Data 0.001 (0.077)	Loss 1.489 (0.859)
Epoch: [8][80/200]	Time 0.260 (0.341)	Data 0.001 (0.076)	Loss 1.127 (0.953)
Epoch: [8][100/200]	Time 0.256 (0.342)	Data 0.001 (0.077)	Loss 0.987 (1.009)
Epoch: [8][120/200]	Time 0.260 (0.340)	Data 0.001 (0.075)	Loss 1.132 (1.042)
Epoch: [8][140/200]	Time 0.264 (0.340)	Data 0.001 (0.075)	Loss 1.193 (1.060)
Epoch: [8][160/200]	Time 0.256 (0.339)	Data 0.001 (0.074)	Loss 1.198 (1.083)
Epoch: [8][180/200]	Time 0.256 (0.339)	Data 0.001 (0.074)	Loss 0.979 (1.095)
Epoch: [8][200/200]	Time 0.260 (0.338)	Data 0.001 (0.073)	Loss 1.019 (1.100)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.133)	Data 0.000 (0.034)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.529709339141846
==> Statistics for epoch 9: 618 clusters
Epoch: [9][20/200]	Time 1.659 (0.377)	Data 1.369 (0.108)	Loss 0.885 (0.205)
Epoch: [9][40/200]	Time 0.258 (0.353)	Data 0.001 (0.086)	Loss 0.946 (0.674)
Epoch: [9][60/200]	Time 0.263 (0.349)	Data 0.001 (0.082)	Loss 1.150 (0.847)
Epoch: [9][80/200]	Time 0.261 (0.346)	Data 0.001 (0.079)	Loss 1.292 (0.940)
Epoch: [9][100/200]	Time 0.255 (0.343)	Data 0.001 (0.077)	Loss 1.052 (0.996)
Epoch: [9][120/200]	Time 0.270 (0.341)	Data 0.001 (0.076)	Loss 1.016 (1.021)
Epoch: [9][140/200]	Time 0.255 (0.340)	Data 0.001 (0.074)	Loss 0.879 (1.037)
Epoch: [9][160/200]	Time 0.261 (0.340)	Data 0.001 (0.074)	Loss 1.144 (1.057)
Epoch: [9][180/200]	Time 0.270 (0.339)	Data 0.001 (0.074)	Loss 1.147 (1.067)
Epoch: [9][200/200]	Time 0.261 (0.339)	Data 0.001 (0.073)	Loss 1.097 (1.081)
Extract Features: [50/76]	Time 0.094 (0.134)	Data 0.000 (0.036)	
Mean AP: 90.9%

 * Finished epoch   9  model mAP: 90.9%  best: 90.9% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.136)	Data 0.000 (0.033)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.313338041305542
==> Statistics for epoch 10: 612 clusters
Epoch: [10][20/200]	Time 1.671 (0.389)	Data 1.373 (0.105)	Loss 0.896 (0.198)
Epoch: [10][40/200]	Time 0.258 (0.363)	Data 0.001 (0.087)	Loss 1.353 (0.666)
Epoch: [10][60/200]	Time 0.268 (0.354)	Data 0.001 (0.082)	Loss 0.967 (0.820)
Epoch: [10][80/200]	Time 0.258 (0.350)	Data 0.000 (0.079)	Loss 1.079 (0.904)
Epoch: [10][100/200]	Time 0.263 (0.346)	Data 0.000 (0.076)	Loss 1.099 (0.938)
Epoch: [10][120/200]	Time 0.259 (0.346)	Data 0.001 (0.076)	Loss 1.559 (0.975)
Epoch: [10][140/200]	Time 0.258 (0.345)	Data 0.001 (0.075)	Loss 0.990 (1.001)
Epoch: [10][160/200]	Time 0.258 (0.344)	Data 0.001 (0.075)	Loss 1.029 (1.002)
Epoch: [10][180/200]	Time 0.260 (0.344)	Data 0.001 (0.075)	Loss 0.943 (1.012)
Epoch: [10][200/200]	Time 0.259 (0.344)	Data 0.000 (0.075)	Loss 1.296 (1.022)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.136)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.23515033721924
==> Statistics for epoch 11: 615 clusters
Epoch: [11][20/200]	Time 1.750 (0.385)	Data 1.426 (0.114)	Loss 1.315 (0.204)
Epoch: [11][40/200]	Time 0.258 (0.359)	Data 0.001 (0.091)	Loss 1.082 (0.602)
Epoch: [11][60/200]	Time 0.262 (0.352)	Data 0.001 (0.086)	Loss 0.845 (0.755)
Epoch: [11][80/200]	Time 0.259 (0.349)	Data 0.001 (0.082)	Loss 1.323 (0.862)
Epoch: [11][100/200]	Time 0.261 (0.347)	Data 0.001 (0.081)	Loss 1.208 (0.905)
Epoch: [11][120/200]	Time 0.257 (0.346)	Data 0.000 (0.080)	Loss 0.735 (0.940)
Epoch: [11][140/200]	Time 0.388 (0.344)	Data 0.000 (0.077)	Loss 1.082 (0.973)
Epoch: [11][160/200]	Time 0.257 (0.344)	Data 0.000 (0.077)	Loss 1.128 (0.994)
Epoch: [11][180/200]	Time 0.262 (0.343)	Data 0.001 (0.077)	Loss 1.030 (1.012)
Epoch: [11][200/200]	Time 0.417 (0.343)	Data 0.001 (0.076)	Loss 1.322 (1.018)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.135)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.206700325012207
==> Statistics for epoch 12: 614 clusters
Epoch: [12][20/200]	Time 1.560 (0.367)	Data 1.262 (0.099)	Loss 0.779 (0.166)
Epoch: [12][40/200]	Time 0.268 (0.353)	Data 0.000 (0.087)	Loss 1.290 (0.589)
Epoch: [12][60/200]	Time 0.260 (0.344)	Data 0.000 (0.079)	Loss 0.838 (0.749)
Epoch: [12][80/200]	Time 0.257 (0.340)	Data 0.000 (0.075)	Loss 0.535 (0.824)
Epoch: [12][100/200]	Time 0.256 (0.338)	Data 0.000 (0.075)	Loss 1.050 (0.887)
Epoch: [12][120/200]	Time 0.255 (0.337)	Data 0.000 (0.074)	Loss 1.041 (0.915)
Epoch: [12][140/200]	Time 0.256 (0.337)	Data 0.000 (0.073)	Loss 1.160 (0.943)
Epoch: [12][160/200]	Time 0.258 (0.336)	Data 0.000 (0.072)	Loss 1.063 (0.958)
Epoch: [12][180/200]	Time 0.257 (0.336)	Data 0.000 (0.072)	Loss 0.976 (0.969)
Epoch: [12][200/200]	Time 0.257 (0.336)	Data 0.001 (0.072)	Loss 0.829 (0.970)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.092 (0.134)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.611284732818604
==> Statistics for epoch 13: 613 clusters
Epoch: [13][20/200]	Time 1.592 (0.367)	Data 1.305 (0.103)	Loss 1.057 (0.186)
Epoch: [13][40/200]	Time 0.269 (0.352)	Data 0.001 (0.087)	Loss 1.206 (0.604)
Epoch: [13][60/200]	Time 0.266 (0.349)	Data 0.001 (0.083)	Loss 1.007 (0.744)
Epoch: [13][80/200]	Time 0.260 (0.345)	Data 0.001 (0.080)	Loss 1.059 (0.821)
Epoch: [13][100/200]	Time 0.257 (0.342)	Data 0.001 (0.076)	Loss 1.138 (0.864)
Epoch: [13][120/200]	Time 0.254 (0.342)	Data 0.001 (0.076)	Loss 1.004 (0.907)
Epoch: [13][140/200]	Time 0.264 (0.341)	Data 0.001 (0.074)	Loss 0.872 (0.931)
Epoch: [13][160/200]	Time 0.257 (0.340)	Data 0.001 (0.074)	Loss 0.913 (0.942)
Epoch: [13][180/200]	Time 0.264 (0.339)	Data 0.001 (0.073)	Loss 1.202 (0.952)
Epoch: [13][200/200]	Time 0.259 (0.338)	Data 0.001 (0.072)	Loss 1.278 (0.960)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.173 (0.135)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.70716166496277
==> Statistics for epoch 14: 615 clusters
Epoch: [14][20/200]	Time 1.740 (0.375)	Data 1.425 (0.104)	Loss 0.901 (0.183)
Epoch: [14][40/200]	Time 0.264 (0.357)	Data 0.001 (0.088)	Loss 1.019 (0.593)
Epoch: [14][60/200]	Time 0.264 (0.349)	Data 0.000 (0.081)	Loss 1.206 (0.717)
Epoch: [14][80/200]	Time 0.260 (0.347)	Data 0.001 (0.080)	Loss 1.020 (0.803)
Epoch: [14][100/200]	Time 0.256 (0.343)	Data 0.000 (0.076)	Loss 1.165 (0.846)
Epoch: [14][120/200]	Time 0.260 (0.342)	Data 0.000 (0.076)	Loss 0.955 (0.872)
Epoch: [14][140/200]	Time 0.397 (0.342)	Data 0.000 (0.076)	Loss 1.254 (0.901)
Epoch: [14][160/200]	Time 0.257 (0.340)	Data 0.000 (0.074)	Loss 0.768 (0.916)
Epoch: [14][180/200]	Time 0.260 (0.340)	Data 0.001 (0.074)	Loss 1.017 (0.929)
Epoch: [14][200/200]	Time 0.257 (0.339)	Data 0.001 (0.074)	Loss 1.047 (0.937)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.131)	Data 0.000 (0.033)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.435604572296143
==> Statistics for epoch 15: 613 clusters
Epoch: [15][20/200]	Time 1.785 (0.370)	Data 1.493 (0.108)	Loss 0.677 (0.155)
Epoch: [15][40/200]	Time 0.260 (0.351)	Data 0.001 (0.090)	Loss 1.154 (0.558)
Epoch: [15][60/200]	Time 0.258 (0.348)	Data 0.001 (0.086)	Loss 1.312 (0.696)
Epoch: [15][80/200]	Time 0.260 (0.345)	Data 0.001 (0.082)	Loss 1.060 (0.763)
Epoch: [15][100/200]	Time 0.271 (0.342)	Data 0.001 (0.078)	Loss 1.176 (0.813)
Epoch: [15][120/200]	Time 0.257 (0.342)	Data 0.001 (0.078)	Loss 0.968 (0.845)
Epoch: [15][140/200]	Time 0.257 (0.341)	Data 0.000 (0.077)	Loss 0.827 (0.869)
Epoch: [15][160/200]	Time 0.258 (0.341)	Data 0.001 (0.077)	Loss 1.069 (0.890)
Epoch: [15][180/200]	Time 0.260 (0.341)	Data 0.000 (0.076)	Loss 1.005 (0.901)
Epoch: [15][200/200]	Time 0.257 (0.341)	Data 0.001 (0.076)	Loss 0.895 (0.911)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.137)	Data 0.000 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.12726354598999
==> Statistics for epoch 16: 612 clusters
Epoch: [16][20/200]	Time 1.831 (0.383)	Data 1.563 (0.116)	Loss 0.955 (0.172)
Epoch: [16][40/200]	Time 0.255 (0.359)	Data 0.001 (0.094)	Loss 0.986 (0.562)
Epoch: [16][60/200]	Time 0.262 (0.350)	Data 0.001 (0.086)	Loss 1.000 (0.690)
Epoch: [16][80/200]	Time 0.259 (0.347)	Data 0.001 (0.083)	Loss 0.818 (0.761)
Epoch: [16][100/200]	Time 0.258 (0.345)	Data 0.001 (0.080)	Loss 1.156 (0.811)
Epoch: [16][120/200]	Time 0.260 (0.342)	Data 0.001 (0.078)	Loss 0.962 (0.824)
Epoch: [16][140/200]	Time 0.259 (0.341)	Data 0.001 (0.077)	Loss 1.032 (0.844)
Epoch: [16][160/200]	Time 0.259 (0.341)	Data 0.001 (0.076)	Loss 0.992 (0.863)
Epoch: [16][180/200]	Time 0.260 (0.340)	Data 0.001 (0.076)	Loss 0.960 (0.872)
Epoch: [16][200/200]	Time 0.258 (0.340)	Data 0.001 (0.075)	Loss 1.141 (0.881)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.142)	Data 0.000 (0.044)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.62503671646118
==> Statistics for epoch 17: 610 clusters
Epoch: [17][20/200]	Time 1.561 (0.372)	Data 1.274 (0.103)	Loss 0.859 (0.173)
Epoch: [17][40/200]	Time 0.373 (0.351)	Data 0.000 (0.085)	Loss 0.877 (0.545)
Epoch: [17][60/200]	Time 0.259 (0.345)	Data 0.001 (0.079)	Loss 0.939 (0.680)
Epoch: [17][80/200]	Time 0.257 (0.341)	Data 0.000 (0.077)	Loss 0.922 (0.746)
Epoch: [17][100/200]	Time 0.263 (0.339)	Data 0.000 (0.075)	Loss 0.954 (0.781)
Epoch: [17][120/200]	Time 0.257 (0.338)	Data 0.001 (0.074)	Loss 0.851 (0.805)
Epoch: [17][140/200]	Time 0.258 (0.339)	Data 0.000 (0.074)	Loss 0.715 (0.825)
Epoch: [17][160/200]	Time 0.255 (0.338)	Data 0.000 (0.073)	Loss 0.979 (0.834)
Epoch: [17][180/200]	Time 0.259 (0.337)	Data 0.000 (0.072)	Loss 0.930 (0.845)
Epoch: [17][200/200]	Time 0.264 (0.337)	Data 0.001 (0.072)	Loss 0.911 (0.863)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.101 (0.136)	Data 0.008 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.672845125198364
==> Statistics for epoch 18: 611 clusters
Epoch: [18][20/200]	Time 1.674 (0.375)	Data 1.379 (0.112)	Loss 0.855 (0.158)
Epoch: [18][40/200]	Time 0.262 (0.350)	Data 0.001 (0.089)	Loss 0.764 (0.510)
Epoch: [18][60/200]	Time 0.263 (0.345)	Data 0.001 (0.081)	Loss 1.039 (0.644)
Epoch: [18][80/200]	Time 0.258 (0.344)	Data 0.001 (0.078)	Loss 0.653 (0.719)
Epoch: [18][100/200]	Time 0.257 (0.342)	Data 0.001 (0.076)	Loss 0.802 (0.757)
Epoch: [18][120/200]	Time 0.258 (0.339)	Data 0.001 (0.074)	Loss 1.063 (0.787)
Epoch: [18][140/200]	Time 0.264 (0.339)	Data 0.001 (0.074)	Loss 0.758 (0.814)
Epoch: [18][160/200]	Time 0.264 (0.338)	Data 0.001 (0.072)	Loss 1.015 (0.818)
Epoch: [18][180/200]	Time 0.257 (0.337)	Data 0.000 (0.072)	Loss 1.161 (0.831)
Epoch: [18][200/200]	Time 0.263 (0.336)	Data 0.001 (0.071)	Loss 0.738 (0.832)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.145)	Data 0.000 (0.047)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.097980737686157
==> Statistics for epoch 19: 609 clusters
Epoch: [19][20/200]	Time 1.744 (0.376)	Data 1.446 (0.110)	Loss 1.023 (0.161)
Epoch: [19][40/200]	Time 0.258 (0.355)	Data 0.001 (0.091)	Loss 0.572 (0.520)
Epoch: [19][60/200]	Time 0.263 (0.349)	Data 0.001 (0.084)	Loss 0.774 (0.629)
Epoch: [19][80/200]	Time 0.358 (0.346)	Data 0.001 (0.081)	Loss 0.922 (0.706)
Epoch: [19][100/200]	Time 0.256 (0.344)	Data 0.001 (0.080)	Loss 0.949 (0.741)
Epoch: [19][120/200]	Time 0.257 (0.343)	Data 0.001 (0.079)	Loss 0.793 (0.771)
Epoch: [19][140/200]	Time 0.256 (0.342)	Data 0.001 (0.078)	Loss 0.853 (0.784)
Epoch: [19][160/200]	Time 0.256 (0.342)	Data 0.001 (0.078)	Loss 0.806 (0.795)
Epoch: [19][180/200]	Time 0.257 (0.342)	Data 0.001 (0.077)	Loss 0.809 (0.810)
Epoch: [19][200/200]	Time 0.257 (0.342)	Data 0.001 (0.077)	Loss 0.896 (0.816)
Extract Features: [50/76]	Time 0.267 (0.141)	Data 0.175 (0.044)	
Mean AP: 91.8%

 * Finished epoch  19  model mAP: 91.8%  best: 91.8% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.135)	Data 0.001 (0.036)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.226353406906128
==> Statistics for epoch 20: 611 clusters
Epoch: [20][20/200]	Time 1.547 (0.366)	Data 1.251 (0.100)	Loss 0.690 (0.141)
Epoch: [20][40/200]	Time 0.262 (0.350)	Data 0.001 (0.085)	Loss 1.062 (0.482)
Epoch: [20][60/200]	Time 0.258 (0.347)	Data 0.001 (0.082)	Loss 0.714 (0.595)
Epoch: [20][80/200]	Time 0.263 (0.343)	Data 0.001 (0.077)	Loss 0.871 (0.672)
Epoch: [20][100/200]	Time 0.259 (0.341)	Data 0.001 (0.075)	Loss 1.080 (0.705)
Epoch: [20][120/200]	Time 0.257 (0.341)	Data 0.001 (0.076)	Loss 0.657 (0.734)
Epoch: [20][140/200]	Time 0.266 (0.341)	Data 0.001 (0.075)	Loss 0.742 (0.750)
Epoch: [20][160/200]	Time 0.263 (0.340)	Data 0.001 (0.073)	Loss 0.663 (0.766)
Epoch: [20][180/200]	Time 0.260 (0.339)	Data 0.001 (0.073)	Loss 0.737 (0.781)
Epoch: [20][200/200]	Time 0.259 (0.338)	Data 0.001 (0.072)	Loss 0.967 (0.789)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.134)	Data 0.000 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.654499530792236
==> Statistics for epoch 21: 609 clusters
Epoch: [21][20/200]	Time 1.729 (0.388)	Data 1.428 (0.115)	Loss 0.755 (0.148)
Epoch: [21][40/200]	Time 0.260 (0.362)	Data 0.001 (0.094)	Loss 0.889 (0.510)
Epoch: [21][60/200]	Time 0.265 (0.352)	Data 0.001 (0.085)	Loss 0.826 (0.628)
Epoch: [21][80/200]	Time 0.258 (0.349)	Data 0.001 (0.083)	Loss 0.954 (0.677)
Epoch: [21][100/200]	Time 0.258 (0.348)	Data 0.001 (0.082)	Loss 0.980 (0.716)
Epoch: [21][120/200]	Time 0.256 (0.347)	Data 0.001 (0.081)	Loss 0.645 (0.741)
Epoch: [21][140/200]	Time 0.258 (0.346)	Data 0.001 (0.080)	Loss 1.045 (0.768)
Epoch: [21][160/200]	Time 0.259 (0.345)	Data 0.001 (0.080)	Loss 1.127 (0.782)
Epoch: [21][180/200]	Time 0.364 (0.344)	Data 0.001 (0.079)	Loss 0.775 (0.784)
Epoch: [21][200/200]	Time 0.258 (0.344)	Data 0.001 (0.078)	Loss 1.009 (0.789)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.136)	Data 0.000 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.417700052261353
==> Statistics for epoch 22: 611 clusters
Epoch: [22][20/200]	Time 1.697 (0.379)	Data 1.396 (0.107)	Loss 0.863 (0.137)
Epoch: [22][40/200]	Time 0.258 (0.360)	Data 0.001 (0.091)	Loss 0.842 (0.459)
Epoch: [22][60/200]	Time 0.271 (0.351)	Data 0.001 (0.083)	Loss 0.766 (0.596)
Epoch: [22][80/200]	Time 0.260 (0.348)	Data 0.001 (0.080)	Loss 0.828 (0.654)
Epoch: [22][100/200]	Time 0.261 (0.346)	Data 0.001 (0.079)	Loss 0.917 (0.697)
Epoch: [22][120/200]	Time 0.258 (0.345)	Data 0.001 (0.078)	Loss 0.962 (0.732)
Epoch: [22][140/200]	Time 0.258 (0.345)	Data 0.001 (0.078)	Loss 0.682 (0.746)
Epoch: [22][160/200]	Time 0.259 (0.345)	Data 0.001 (0.078)	Loss 0.700 (0.756)
Epoch: [22][180/200]	Time 0.260 (0.344)	Data 0.001 (0.077)	Loss 1.096 (0.767)
Epoch: [22][200/200]	Time 0.260 (0.344)	Data 0.001 (0.077)	Loss 1.227 (0.782)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.142)	Data 0.000 (0.044)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.558632135391235
==> Statistics for epoch 23: 611 clusters
Epoch: [23][20/200]	Time 1.809 (0.383)	Data 1.520 (0.115)	Loss 1.199 (0.168)
Epoch: [23][40/200]	Time 0.265 (0.362)	Data 0.001 (0.096)	Loss 0.865 (0.504)
Epoch: [23][60/200]	Time 0.259 (0.356)	Data 0.001 (0.090)	Loss 0.677 (0.632)
Epoch: [23][80/200]	Time 0.259 (0.351)	Data 0.001 (0.085)	Loss 0.738 (0.690)
Epoch: [23][100/200]	Time 0.264 (0.348)	Data 0.000 (0.082)	Loss 1.078 (0.724)
Epoch: [23][120/200]	Time 0.262 (0.346)	Data 0.000 (0.080)	Loss 0.647 (0.748)
Epoch: [23][140/200]	Time 0.256 (0.346)	Data 0.001 (0.080)	Loss 0.939 (0.765)
Epoch: [23][160/200]	Time 0.258 (0.344)	Data 0.001 (0.079)	Loss 0.761 (0.766)
Epoch: [23][180/200]	Time 0.260 (0.344)	Data 0.001 (0.079)	Loss 0.592 (0.776)
Epoch: [23][200/200]	Time 0.262 (0.344)	Data 0.001 (0.079)	Loss 0.654 (0.784)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.101 (0.138)	Data 0.007 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.190653562545776
==> Statistics for epoch 24: 611 clusters
Epoch: [24][20/200]	Time 1.683 (0.376)	Data 1.405 (0.111)	Loss 0.714 (0.140)
Epoch: [24][40/200]	Time 0.258 (0.356)	Data 0.001 (0.093)	Loss 0.712 (0.467)
Epoch: [24][60/200]	Time 0.276 (0.348)	Data 0.001 (0.084)	Loss 1.089 (0.599)
Epoch: [24][80/200]	Time 0.258 (0.344)	Data 0.001 (0.078)	Loss 0.601 (0.642)
Epoch: [24][100/200]	Time 0.261 (0.341)	Data 0.001 (0.076)	Loss 0.824 (0.665)
Epoch: [24][120/200]	Time 0.261 (0.340)	Data 0.004 (0.075)	Loss 0.870 (0.684)
Epoch: [24][140/200]	Time 0.257 (0.338)	Data 0.001 (0.073)	Loss 0.933 (0.705)
Epoch: [24][160/200]	Time 0.264 (0.338)	Data 0.001 (0.073)	Loss 0.963 (0.726)
Epoch: [24][180/200]	Time 0.263 (0.338)	Data 0.001 (0.073)	Loss 0.922 (0.739)
Epoch: [24][200/200]	Time 0.257 (0.338)	Data 0.001 (0.073)	Loss 0.733 (0.749)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.132)	Data 0.000 (0.033)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.212523698806763
==> Statistics for epoch 25: 610 clusters
Epoch: [25][20/200]	Time 1.620 (0.371)	Data 1.331 (0.105)	Loss 0.657 (0.132)
Epoch: [25][40/200]	Time 0.256 (0.347)	Data 0.001 (0.085)	Loss 0.798 (0.470)
Epoch: [25][60/200]	Time 0.256 (0.342)	Data 0.001 (0.079)	Loss 0.893 (0.592)
Epoch: [25][80/200]	Time 0.263 (0.340)	Data 0.001 (0.076)	Loss 0.911 (0.645)
Epoch: [25][100/200]	Time 0.267 (0.337)	Data 0.001 (0.073)	Loss 0.861 (0.689)
Epoch: [25][120/200]	Time 0.260 (0.336)	Data 0.001 (0.072)	Loss 0.884 (0.713)
Epoch: [25][140/200]	Time 0.261 (0.335)	Data 0.001 (0.070)	Loss 0.884 (0.730)
Epoch: [25][160/200]	Time 0.261 (0.334)	Data 0.001 (0.070)	Loss 0.998 (0.740)
Epoch: [25][180/200]	Time 0.266 (0.334)	Data 0.001 (0.069)	Loss 0.579 (0.751)
Epoch: [25][200/200]	Time 0.260 (0.333)	Data 0.001 (0.069)	Loss 0.904 (0.768)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.141)	Data 0.000 (0.046)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.96362018585205
==> Statistics for epoch 26: 610 clusters
Epoch: [26][20/200]	Time 1.772 (0.375)	Data 1.478 (0.106)	Loss 0.696 (0.139)
Epoch: [26][40/200]	Time 0.261 (0.353)	Data 0.001 (0.086)	Loss 0.914 (0.477)
Epoch: [26][60/200]	Time 0.270 (0.345)	Data 0.001 (0.078)	Loss 0.949 (0.597)
Epoch: [26][80/200]	Time 0.257 (0.341)	Data 0.001 (0.074)	Loss 0.775 (0.661)
Epoch: [26][100/200]	Time 0.258 (0.339)	Data 0.001 (0.072)	Loss 0.824 (0.693)
Epoch: [26][120/200]	Time 0.271 (0.338)	Data 0.001 (0.072)	Loss 0.781 (0.714)
Epoch: [26][140/200]	Time 0.261 (0.338)	Data 0.001 (0.072)	Loss 0.631 (0.731)
Epoch: [26][160/200]	Time 0.265 (0.338)	Data 0.001 (0.072)	Loss 0.800 (0.744)
Epoch: [26][180/200]	Time 0.261 (0.338)	Data 0.001 (0.072)	Loss 0.726 (0.751)
Epoch: [26][200/200]	Time 0.262 (0.338)	Data 0.001 (0.071)	Loss 0.942 (0.763)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.092 (0.138)	Data 0.000 (0.040)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.53096580505371
==> Statistics for epoch 27: 610 clusters
Epoch: [27][20/200]	Time 1.571 (0.368)	Data 1.282 (0.101)	Loss 0.750 (0.148)
Epoch: [27][40/200]	Time 0.269 (0.346)	Data 0.001 (0.083)	Loss 0.885 (0.500)
Epoch: [27][60/200]	Time 0.266 (0.341)	Data 0.001 (0.076)	Loss 0.515 (0.594)
Epoch: [27][80/200]	Time 0.257 (0.339)	Data 0.001 (0.074)	Loss 0.981 (0.649)
Epoch: [27][100/200]	Time 0.275 (0.338)	Data 0.001 (0.073)	Loss 0.883 (0.676)
Epoch: [27][120/200]	Time 0.260 (0.337)	Data 0.001 (0.072)	Loss 1.101 (0.695)
Epoch: [27][140/200]	Time 0.255 (0.335)	Data 0.001 (0.071)	Loss 0.711 (0.719)
Epoch: [27][160/200]	Time 0.257 (0.334)	Data 0.000 (0.070)	Loss 0.840 (0.735)
Epoch: [27][180/200]	Time 0.261 (0.334)	Data 0.001 (0.070)	Loss 0.755 (0.741)
Epoch: [27][200/200]	Time 0.259 (0.334)	Data 0.001 (0.070)	Loss 0.801 (0.756)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.092 (0.133)	Data 0.000 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.38787817955017
==> Statistics for epoch 28: 611 clusters
Epoch: [28][20/200]	Time 1.598 (0.368)	Data 1.301 (0.107)	Loss 0.848 (0.151)
Epoch: [28][40/200]	Time 0.258 (0.352)	Data 0.001 (0.089)	Loss 0.861 (0.498)
Epoch: [28][60/200]	Time 0.264 (0.348)	Data 0.001 (0.084)	Loss 1.034 (0.618)
Epoch: [28][80/200]	Time 0.264 (0.345)	Data 0.001 (0.081)	Loss 1.057 (0.679)
Epoch: [28][100/200]	Time 0.257 (0.344)	Data 0.001 (0.080)	Loss 0.826 (0.711)
Epoch: [28][120/200]	Time 0.258 (0.342)	Data 0.001 (0.078)	Loss 0.947 (0.739)
Epoch: [28][140/200]	Time 0.255 (0.339)	Data 0.001 (0.076)	Loss 1.219 (0.753)
Epoch: [28][160/200]	Time 0.258 (0.338)	Data 0.001 (0.075)	Loss 1.168 (0.762)
Epoch: [28][180/200]	Time 0.256 (0.337)	Data 0.001 (0.073)	Loss 0.878 (0.778)
Epoch: [28][200/200]	Time 0.259 (0.337)	Data 0.001 (0.073)	Loss 0.904 (0.780)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.126 (0.136)	Data 0.029 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.514359951019287
==> Statistics for epoch 29: 612 clusters
Epoch: [29][20/200]	Time 1.610 (0.367)	Data 1.327 (0.104)	Loss 0.808 (0.142)
Epoch: [29][40/200]	Time 0.260 (0.348)	Data 0.001 (0.085)	Loss 0.805 (0.476)
Epoch: [29][60/200]	Time 0.257 (0.342)	Data 0.000 (0.078)	Loss 0.762 (0.605)
Epoch: [29][80/200]	Time 0.259 (0.340)	Data 0.000 (0.075)	Loss 0.847 (0.672)
Epoch: [29][100/200]	Time 0.262 (0.338)	Data 0.001 (0.072)	Loss 1.012 (0.704)
Epoch: [29][120/200]	Time 0.263 (0.336)	Data 0.001 (0.071)	Loss 0.762 (0.722)
Epoch: [29][140/200]	Time 0.271 (0.336)	Data 0.001 (0.071)	Loss 1.117 (0.740)
Epoch: [29][160/200]	Time 0.263 (0.336)	Data 0.001 (0.070)	Loss 0.895 (0.749)
Epoch: [29][180/200]	Time 0.257 (0.335)	Data 0.000 (0.069)	Loss 1.057 (0.766)
Epoch: [29][200/200]	Time 0.263 (0.335)	Data 0.000 (0.069)	Loss 0.748 (0.770)
Extract Features: [50/76]	Time 0.094 (0.144)	Data 0.000 (0.046)	
Mean AP: 92.1%

 * Finished epoch  29  model mAP: 92.1%  best: 92.1% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.092 (0.139)	Data 0.000 (0.040)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.528892517089844
==> Statistics for epoch 30: 611 clusters
Epoch: [30][20/200]	Time 1.641 (0.372)	Data 1.339 (0.103)	Loss 0.787 (0.151)
Epoch: [30][40/200]	Time 0.255 (0.349)	Data 0.001 (0.083)	Loss 0.519 (0.497)
Epoch: [30][60/200]	Time 0.262 (0.341)	Data 0.001 (0.076)	Loss 0.718 (0.603)
Epoch: [30][80/200]	Time 0.258 (0.337)	Data 0.000 (0.072)	Loss 0.859 (0.654)
Epoch: [30][100/200]	Time 0.257 (0.336)	Data 0.001 (0.071)	Loss 0.700 (0.682)
Epoch: [30][120/200]	Time 0.256 (0.333)	Data 0.001 (0.069)	Loss 0.414 (0.702)
Epoch: [30][140/200]	Time 0.255 (0.333)	Data 0.000 (0.068)	Loss 0.862 (0.721)
Epoch: [30][160/200]	Time 0.258 (0.331)	Data 0.001 (0.067)	Loss 0.816 (0.729)
Epoch: [30][180/200]	Time 0.259 (0.331)	Data 0.001 (0.067)	Loss 0.750 (0.741)
Epoch: [30][200/200]	Time 0.259 (0.331)	Data 0.001 (0.066)	Loss 0.956 (0.755)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.182 (0.134)	Data 0.009 (0.035)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.663322687149048
==> Statistics for epoch 31: 610 clusters
Epoch: [31][20/200]	Time 1.682 (0.371)	Data 1.378 (0.107)	Loss 0.682 (0.136)
Epoch: [31][40/200]	Time 0.257 (0.355)	Data 0.001 (0.091)	Loss 1.000 (0.471)
Epoch: [31][60/200]	Time 0.263 (0.346)	Data 0.001 (0.081)	Loss 0.858 (0.571)
Epoch: [31][80/200]	Time 0.260 (0.340)	Data 0.001 (0.076)	Loss 0.949 (0.632)
Epoch: [31][100/200]	Time 0.259 (0.338)	Data 0.001 (0.074)	Loss 0.834 (0.672)
Epoch: [31][120/200]	Time 0.258 (0.338)	Data 0.001 (0.074)	Loss 0.583 (0.696)
Epoch: [31][140/200]	Time 0.260 (0.337)	Data 0.001 (0.073)	Loss 0.534 (0.724)
Epoch: [31][160/200]	Time 0.257 (0.336)	Data 0.001 (0.072)	Loss 0.806 (0.735)
Epoch: [31][180/200]	Time 0.258 (0.336)	Data 0.001 (0.071)	Loss 1.098 (0.745)
Epoch: [31][200/200]	Time 0.258 (0.336)	Data 0.001 (0.071)	Loss 1.130 (0.754)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.143 (0.138)	Data 0.051 (0.041)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.490511655807495
==> Statistics for epoch 32: 609 clusters
Epoch: [32][20/200]	Time 1.591 (0.372)	Data 1.302 (0.105)	Loss 0.940 (0.147)
Epoch: [32][40/200]	Time 0.266 (0.351)	Data 0.001 (0.085)	Loss 0.613 (0.459)
Epoch: [32][60/200]	Time 0.260 (0.344)	Data 0.000 (0.078)	Loss 0.575 (0.585)
Epoch: [32][80/200]	Time 0.260 (0.342)	Data 0.000 (0.075)	Loss 0.781 (0.652)
Epoch: [32][100/200]	Time 0.264 (0.341)	Data 0.001 (0.076)	Loss 0.799 (0.683)
Epoch: [32][120/200]	Time 0.262 (0.340)	Data 0.002 (0.074)	Loss 0.634 (0.710)
Epoch: [32][140/200]	Time 0.260 (0.340)	Data 0.001 (0.074)	Loss 0.850 (0.726)
Epoch: [32][160/200]	Time 0.272 (0.341)	Data 0.001 (0.074)	Loss 0.592 (0.739)
Epoch: [32][180/200]	Time 0.270 (0.340)	Data 0.001 (0.073)	Loss 0.832 (0.754)
Epoch: [32][200/200]	Time 0.260 (0.339)	Data 0.001 (0.072)	Loss 1.243 (0.765)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.136)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.406389713287354
==> Statistics for epoch 33: 610 clusters
Epoch: [33][20/200]	Time 1.550 (0.368)	Data 1.255 (0.101)	Loss 0.974 (0.153)
Epoch: [33][40/200]	Time 0.265 (0.349)	Data 0.001 (0.082)	Loss 0.806 (0.494)
Epoch: [33][60/200]	Time 0.259 (0.342)	Data 0.001 (0.074)	Loss 1.083 (0.603)
Epoch: [33][80/200]	Time 0.259 (0.339)	Data 0.001 (0.072)	Loss 1.021 (0.661)
Epoch: [33][100/200]	Time 0.257 (0.338)	Data 0.001 (0.070)	Loss 0.647 (0.683)
Epoch: [33][120/200]	Time 0.260 (0.336)	Data 0.001 (0.068)	Loss 0.702 (0.701)
Epoch: [33][140/200]	Time 0.258 (0.335)	Data 0.000 (0.068)	Loss 1.098 (0.722)
Epoch: [33][160/200]	Time 0.259 (0.334)	Data 0.001 (0.067)	Loss 0.703 (0.733)
Epoch: [33][180/200]	Time 0.275 (0.333)	Data 0.001 (0.067)	Loss 0.781 (0.740)
Epoch: [33][200/200]	Time 0.259 (0.333)	Data 0.001 (0.067)	Loss 0.661 (0.741)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.130)	Data 0.000 (0.033)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.944037199020386
==> Statistics for epoch 34: 611 clusters
Epoch: [34][20/200]	Time 1.541 (0.367)	Data 1.254 (0.100)	Loss 0.672 (0.129)
Epoch: [34][40/200]	Time 0.263 (0.349)	Data 0.001 (0.083)	Loss 1.011 (0.477)
Epoch: [34][60/200]	Time 0.261 (0.344)	Data 0.001 (0.077)	Loss 1.045 (0.600)
Epoch: [34][80/200]	Time 0.258 (0.340)	Data 0.001 (0.074)	Loss 1.010 (0.682)
Epoch: [34][100/200]	Time 0.261 (0.336)	Data 0.001 (0.070)	Loss 0.935 (0.703)
Epoch: [34][120/200]	Time 0.264 (0.335)	Data 0.001 (0.070)	Loss 0.747 (0.718)
Epoch: [34][140/200]	Time 0.266 (0.333)	Data 0.001 (0.068)	Loss 0.746 (0.736)
Epoch: [34][160/200]	Time 0.265 (0.333)	Data 0.001 (0.068)	Loss 0.655 (0.748)
Epoch: [34][180/200]	Time 0.260 (0.334)	Data 0.001 (0.068)	Loss 0.733 (0.751)
Epoch: [34][200/200]	Time 0.257 (0.333)	Data 0.001 (0.068)	Loss 0.775 (0.756)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.180 (0.131)	Data 0.087 (0.034)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.849454164505005
==> Statistics for epoch 35: 610 clusters
Epoch: [35][20/200]	Time 1.817 (0.372)	Data 1.408 (0.103)	Loss 0.970 (0.147)
Epoch: [35][40/200]	Time 0.268 (0.353)	Data 0.000 (0.089)	Loss 0.887 (0.455)
Epoch: [35][60/200]	Time 0.260 (0.348)	Data 0.000 (0.083)	Loss 1.100 (0.573)
Epoch: [35][80/200]	Time 0.261 (0.343)	Data 0.001 (0.078)	Loss 1.014 (0.644)
Epoch: [35][100/200]	Time 0.257 (0.342)	Data 0.001 (0.076)	Loss 0.799 (0.672)
Epoch: [35][120/200]	Time 0.256 (0.341)	Data 0.001 (0.075)	Loss 0.534 (0.692)
Epoch: [35][140/200]	Time 0.272 (0.339)	Data 0.001 (0.073)	Loss 0.913 (0.718)
Epoch: [35][160/200]	Time 0.259 (0.339)	Data 0.001 (0.072)	Loss 0.781 (0.727)
Epoch: [35][180/200]	Time 0.263 (0.339)	Data 0.001 (0.073)	Loss 0.784 (0.736)
Epoch: [35][200/200]	Time 0.258 (0.338)	Data 0.001 (0.071)	Loss 0.804 (0.743)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.093 (0.136)	Data 0.000 (0.040)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.685068130493164
==> Statistics for epoch 36: 610 clusters
Epoch: [36][20/200]	Time 1.696 (0.372)	Data 1.380 (0.107)	Loss 0.958 (0.143)
Epoch: [36][40/200]	Time 0.361 (0.352)	Data 0.001 (0.087)	Loss 0.772 (0.435)
Epoch: [36][60/200]	Time 0.257 (0.346)	Data 0.001 (0.083)	Loss 0.634 (0.567)
Epoch: [36][80/200]	Time 0.385 (0.346)	Data 0.001 (0.081)	Loss 0.914 (0.632)
Epoch: [36][100/200]	Time 0.258 (0.343)	Data 0.001 (0.078)	Loss 0.711 (0.668)
Epoch: [36][120/200]	Time 0.261 (0.341)	Data 0.001 (0.077)	Loss 1.293 (0.692)
Epoch: [36][140/200]	Time 0.276 (0.341)	Data 0.001 (0.076)	Loss 0.545 (0.704)
Epoch: [36][160/200]	Time 0.256 (0.341)	Data 0.001 (0.076)	Loss 0.960 (0.714)
Epoch: [36][180/200]	Time 0.260 (0.340)	Data 0.001 (0.075)	Loss 0.787 (0.725)
Epoch: [36][200/200]	Time 0.260 (0.340)	Data 0.001 (0.075)	Loss 0.861 (0.728)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.139)	Data 0.000 (0.040)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.887545824050903
==> Statistics for epoch 37: 611 clusters
Epoch: [37][20/200]	Time 1.636 (0.374)	Data 1.360 (0.109)	Loss 0.765 (0.137)
Epoch: [37][40/200]	Time 0.262 (0.352)	Data 0.000 (0.087)	Loss 0.895 (0.462)
Epoch: [37][60/200]	Time 0.262 (0.344)	Data 0.000 (0.080)	Loss 0.861 (0.568)
Epoch: [37][80/200]	Time 0.257 (0.339)	Data 0.001 (0.075)	Loss 0.773 (0.640)
Epoch: [37][100/200]	Time 0.254 (0.337)	Data 0.000 (0.072)	Loss 0.798 (0.669)
Epoch: [37][120/200]	Time 0.258 (0.335)	Data 0.001 (0.071)	Loss 0.967 (0.693)
Epoch: [37][140/200]	Time 0.257 (0.335)	Data 0.001 (0.071)	Loss 0.911 (0.716)
Epoch: [37][160/200]	Time 0.257 (0.335)	Data 0.001 (0.070)	Loss 0.864 (0.733)
Epoch: [37][180/200]	Time 0.258 (0.334)	Data 0.001 (0.069)	Loss 0.531 (0.743)
Epoch: [37][200/200]	Time 0.258 (0.334)	Data 0.001 (0.069)	Loss 0.962 (0.753)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.130 (0.134)	Data 0.038 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.181211709976196
==> Statistics for epoch 38: 612 clusters
Epoch: [38][20/200]	Time 1.620 (0.373)	Data 1.338 (0.106)	Loss 0.706 (0.130)
Epoch: [38][40/200]	Time 0.276 (0.353)	Data 0.001 (0.087)	Loss 0.984 (0.462)
Epoch: [38][60/200]	Time 0.260 (0.342)	Data 0.001 (0.078)	Loss 1.183 (0.578)
Epoch: [38][80/200]	Time 0.256 (0.339)	Data 0.001 (0.075)	Loss 1.000 (0.642)
Epoch: [38][100/200]	Time 0.256 (0.337)	Data 0.001 (0.074)	Loss 0.831 (0.678)
Epoch: [38][120/200]	Time 0.259 (0.336)	Data 0.001 (0.073)	Loss 0.553 (0.697)
Epoch: [38][140/200]	Time 0.258 (0.335)	Data 0.001 (0.071)	Loss 0.976 (0.718)
Epoch: [38][160/200]	Time 0.259 (0.334)	Data 0.001 (0.070)	Loss 0.939 (0.731)
Epoch: [38][180/200]	Time 0.259 (0.334)	Data 0.001 (0.071)	Loss 0.801 (0.738)
Epoch: [38][200/200]	Time 0.274 (0.334)	Data 0.001 (0.070)	Loss 0.629 (0.749)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.175 (0.132)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.03702402114868
==> Statistics for epoch 39: 611 clusters
Epoch: [39][20/200]	Time 1.667 (0.371)	Data 1.374 (0.104)	Loss 0.556 (0.123)
Epoch: [39][40/200]	Time 0.259 (0.345)	Data 0.001 (0.081)	Loss 0.582 (0.436)
Epoch: [39][60/200]	Time 0.263 (0.340)	Data 0.001 (0.074)	Loss 0.660 (0.546)
Epoch: [39][80/200]	Time 0.258 (0.336)	Data 0.001 (0.072)	Loss 0.834 (0.606)
Epoch: [39][100/200]	Time 0.259 (0.334)	Data 0.001 (0.070)	Loss 1.054 (0.654)
Epoch: [39][120/200]	Time 0.258 (0.333)	Data 0.001 (0.069)	Loss 0.913 (0.681)
Epoch: [39][140/200]	Time 0.259 (0.333)	Data 0.001 (0.069)	Loss 0.601 (0.698)
Epoch: [39][160/200]	Time 0.262 (0.333)	Data 0.001 (0.069)	Loss 0.659 (0.711)
Epoch: [39][180/200]	Time 0.261 (0.333)	Data 0.001 (0.069)	Loss 0.668 (0.719)
Epoch: [39][200/200]	Time 0.262 (0.334)	Data 0.001 (0.069)	Loss 0.928 (0.724)
Extract Features: [50/76]	Time 0.093 (0.134)	Data 0.000 (0.036)	
Mean AP: 92.1%

 * Finished epoch  39  model mAP: 92.1%  best: 92.1% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.129)	Data 0.000 (0.030)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.735535621643066
==> Statistics for epoch 40: 611 clusters
Epoch: [40][20/200]	Time 1.585 (0.367)	Data 1.300 (0.101)	Loss 0.697 (0.130)
Epoch: [40][40/200]	Time 0.263 (0.351)	Data 0.001 (0.088)	Loss 0.640 (0.454)
Epoch: [40][60/200]	Time 0.260 (0.345)	Data 0.001 (0.081)	Loss 0.559 (0.575)
Epoch: [40][80/200]	Time 0.270 (0.343)	Data 0.001 (0.078)	Loss 0.864 (0.639)
Epoch: [40][100/200]	Time 0.258 (0.340)	Data 0.001 (0.075)	Loss 0.817 (0.666)
Epoch: [40][120/200]	Time 0.257 (0.339)	Data 0.001 (0.074)	Loss 0.761 (0.698)
Epoch: [40][140/200]	Time 0.257 (0.337)	Data 0.001 (0.072)	Loss 0.503 (0.707)
Epoch: [40][160/200]	Time 0.258 (0.335)	Data 0.002 (0.071)	Loss 0.820 (0.722)
Epoch: [40][180/200]	Time 0.260 (0.335)	Data 0.001 (0.071)	Loss 0.662 (0.729)
Epoch: [40][200/200]	Time 0.262 (0.334)	Data 0.001 (0.070)	Loss 0.683 (0.737)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.131)	Data 0.000 (0.032)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.6453959941864
==> Statistics for epoch 41: 612 clusters
Epoch: [41][20/200]	Time 1.635 (0.363)	Data 1.349 (0.101)	Loss 0.816 (0.136)
Epoch: [41][40/200]	Time 0.259 (0.345)	Data 0.001 (0.081)	Loss 0.871 (0.456)
Epoch: [41][60/200]	Time 0.255 (0.341)	Data 0.001 (0.077)	Loss 0.882 (0.583)
Epoch: [41][80/200]	Time 0.258 (0.340)	Data 0.001 (0.075)	Loss 0.808 (0.632)
Epoch: [41][100/200]	Time 0.260 (0.337)	Data 0.001 (0.072)	Loss 0.498 (0.655)
Epoch: [41][120/200]	Time 0.260 (0.336)	Data 0.001 (0.071)	Loss 0.623 (0.673)
Epoch: [41][140/200]	Time 0.254 (0.336)	Data 0.001 (0.072)	Loss 0.715 (0.693)
Epoch: [41][160/200]	Time 0.261 (0.335)	Data 0.001 (0.071)	Loss 1.056 (0.708)
Epoch: [41][180/200]	Time 0.258 (0.335)	Data 0.001 (0.070)	Loss 0.597 (0.717)
Epoch: [41][200/200]	Time 0.257 (0.335)	Data 0.001 (0.070)	Loss 0.768 (0.729)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.138)	Data 0.000 (0.040)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.15645718574524
==> Statistics for epoch 42: 611 clusters
Epoch: [42][20/200]	Time 1.685 (0.381)	Data 1.384 (0.111)	Loss 1.187 (0.161)
Epoch: [42][40/200]	Time 0.264 (0.355)	Data 0.001 (0.086)	Loss 0.782 (0.464)
Epoch: [42][60/200]	Time 0.259 (0.349)	Data 0.001 (0.082)	Loss 0.988 (0.600)
Epoch: [42][80/200]	Time 0.257 (0.345)	Data 0.000 (0.077)	Loss 0.922 (0.656)
Epoch: [42][100/200]	Time 0.260 (0.343)	Data 0.000 (0.077)	Loss 0.732 (0.679)
Epoch: [42][120/200]	Time 0.261 (0.340)	Data 0.000 (0.075)	Loss 0.775 (0.705)
Epoch: [42][140/200]	Time 0.258 (0.339)	Data 0.000 (0.074)	Loss 0.707 (0.712)
Epoch: [42][160/200]	Time 0.258 (0.337)	Data 0.000 (0.072)	Loss 0.844 (0.719)
Epoch: [42][180/200]	Time 0.259 (0.338)	Data 0.000 (0.073)	Loss 1.223 (0.730)
Epoch: [42][200/200]	Time 0.259 (0.337)	Data 0.000 (0.072)	Loss 0.712 (0.736)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.130)	Data 0.000 (0.032)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.13204050064087
==> Statistics for epoch 43: 610 clusters
Epoch: [43][20/200]	Time 1.675 (0.366)	Data 1.371 (0.104)	Loss 0.740 (0.130)
Epoch: [43][40/200]	Time 0.259 (0.348)	Data 0.001 (0.084)	Loss 1.016 (0.443)
Epoch: [43][60/200]	Time 0.259 (0.345)	Data 0.001 (0.081)	Loss 0.613 (0.578)
Epoch: [43][80/200]	Time 0.259 (0.343)	Data 0.001 (0.079)	Loss 0.658 (0.634)
Epoch: [43][100/200]	Time 0.259 (0.341)	Data 0.001 (0.076)	Loss 0.816 (0.656)
Epoch: [43][120/200]	Time 0.258 (0.340)	Data 0.001 (0.075)	Loss 0.849 (0.695)
Epoch: [43][140/200]	Time 0.264 (0.339)	Data 0.001 (0.074)	Loss 0.901 (0.707)
Epoch: [43][160/200]	Time 0.270 (0.338)	Data 0.001 (0.072)	Loss 0.738 (0.716)
Epoch: [43][180/200]	Time 0.258 (0.337)	Data 0.001 (0.071)	Loss 1.026 (0.728)
Epoch: [43][200/200]	Time 0.259 (0.336)	Data 0.001 (0.071)	Loss 0.801 (0.735)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.138)	Data 0.000 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.354661226272583
==> Statistics for epoch 44: 610 clusters
Epoch: [44][20/200]	Time 1.592 (0.374)	Data 1.293 (0.105)	Loss 0.725 (0.138)
Epoch: [44][40/200]	Time 0.257 (0.350)	Data 0.001 (0.083)	Loss 0.936 (0.470)
Epoch: [44][60/200]	Time 0.264 (0.342)	Data 0.001 (0.077)	Loss 0.733 (0.578)
Epoch: [44][80/200]	Time 0.259 (0.340)	Data 0.001 (0.075)	Loss 0.829 (0.640)
Epoch: [44][100/200]	Time 0.255 (0.338)	Data 0.001 (0.073)	Loss 0.817 (0.673)
Epoch: [44][120/200]	Time 0.255 (0.336)	Data 0.001 (0.071)	Loss 0.782 (0.687)
Epoch: [44][140/200]	Time 0.257 (0.335)	Data 0.001 (0.071)	Loss 0.815 (0.704)
Epoch: [44][160/200]	Time 0.273 (0.334)	Data 0.001 (0.070)	Loss 0.663 (0.717)
Epoch: [44][180/200]	Time 0.256 (0.333)	Data 0.001 (0.069)	Loss 0.970 (0.724)
Epoch: [44][200/200]	Time 0.374 (0.334)	Data 0.001 (0.069)	Loss 0.826 (0.737)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.149 (0.135)	Data 0.055 (0.039)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.510707139968872
==> Statistics for epoch 45: 611 clusters
Epoch: [45][20/200]	Time 1.682 (0.374)	Data 1.396 (0.105)	Loss 0.674 (0.129)
Epoch: [45][40/200]	Time 0.270 (0.357)	Data 0.001 (0.089)	Loss 0.699 (0.466)
Epoch: [45][60/200]	Time 0.261 (0.346)	Data 0.001 (0.080)	Loss 0.883 (0.586)
Epoch: [45][80/200]	Time 0.260 (0.341)	Data 0.001 (0.075)	Loss 0.822 (0.655)
Epoch: [45][100/200]	Time 0.260 (0.338)	Data 0.001 (0.072)	Loss 0.773 (0.680)
Epoch: [45][120/200]	Time 0.254 (0.336)	Data 0.000 (0.070)	Loss 0.790 (0.703)
Epoch: [45][140/200]	Time 0.257 (0.334)	Data 0.001 (0.069)	Loss 0.687 (0.709)
Epoch: [45][160/200]	Time 0.258 (0.334)	Data 0.001 (0.069)	Loss 1.053 (0.724)
Epoch: [45][180/200]	Time 0.260 (0.333)	Data 0.001 (0.068)	Loss 0.908 (0.736)
Epoch: [45][200/200]	Time 0.259 (0.332)	Data 0.001 (0.068)	Loss 0.806 (0.741)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.134)	Data 0.000 (0.037)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.69158387184143
==> Statistics for epoch 46: 610 clusters
Epoch: [46][20/200]	Time 1.737 (0.371)	Data 1.464 (0.107)	Loss 0.595 (0.120)
Epoch: [46][40/200]	Time 0.257 (0.348)	Data 0.001 (0.084)	Loss 1.085 (0.445)
Epoch: [46][60/200]	Time 0.258 (0.343)	Data 0.001 (0.081)	Loss 0.706 (0.556)
Epoch: [46][80/200]	Time 0.267 (0.342)	Data 0.002 (0.078)	Loss 0.769 (0.613)
Epoch: [46][100/200]	Time 0.258 (0.340)	Data 0.001 (0.077)	Loss 0.838 (0.649)
Epoch: [46][120/200]	Time 0.258 (0.339)	Data 0.001 (0.075)	Loss 0.580 (0.675)
Epoch: [46][140/200]	Time 0.256 (0.339)	Data 0.001 (0.074)	Loss 0.818 (0.686)
Epoch: [46][160/200]	Time 0.258 (0.337)	Data 0.001 (0.073)	Loss 0.810 (0.705)
Epoch: [46][180/200]	Time 0.264 (0.338)	Data 0.001 (0.073)	Loss 0.679 (0.718)
Epoch: [46][200/200]	Time 0.259 (0.338)	Data 0.001 (0.073)	Loss 0.656 (0.729)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.140)	Data 0.000 (0.044)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.413758516311646
==> Statistics for epoch 47: 611 clusters
Epoch: [47][20/200]	Time 1.703 (0.371)	Data 1.414 (0.109)	Loss 0.857 (0.135)
Epoch: [47][40/200]	Time 0.258 (0.349)	Data 0.001 (0.085)	Loss 1.034 (0.473)
Epoch: [47][60/200]	Time 0.262 (0.344)	Data 0.001 (0.080)	Loss 0.752 (0.547)
Epoch: [47][80/200]	Time 0.261 (0.342)	Data 0.001 (0.078)	Loss 0.656 (0.614)
Epoch: [47][100/200]	Time 0.263 (0.343)	Data 0.001 (0.077)	Loss 0.787 (0.637)
Epoch: [47][120/200]	Time 0.260 (0.341)	Data 0.000 (0.075)	Loss 0.850 (0.662)
Epoch: [47][140/200]	Time 0.262 (0.340)	Data 0.001 (0.074)	Loss 0.890 (0.677)
Epoch: [47][160/200]	Time 0.258 (0.341)	Data 0.001 (0.074)	Loss 0.682 (0.691)
Epoch: [47][180/200]	Time 0.259 (0.339)	Data 0.001 (0.073)	Loss 0.774 (0.698)
Epoch: [47][200/200]	Time 0.259 (0.339)	Data 0.001 (0.073)	Loss 0.953 (0.712)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.094 (0.137)	Data 0.000 (0.038)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.647951126098633
==> Statistics for epoch 48: 611 clusters
Epoch: [48][20/200]	Time 1.752 (0.381)	Data 1.479 (0.112)	Loss 0.847 (0.143)
Epoch: [48][40/200]	Time 0.258 (0.356)	Data 0.001 (0.089)	Loss 0.812 (0.467)
Epoch: [48][60/200]	Time 0.291 (0.347)	Data 0.001 (0.080)	Loss 0.796 (0.580)
Epoch: [48][80/200]	Time 0.261 (0.345)	Data 0.001 (0.078)	Loss 0.753 (0.641)
Epoch: [48][100/200]	Time 0.260 (0.343)	Data 0.001 (0.076)	Loss 0.682 (0.673)
Epoch: [48][120/200]	Time 0.255 (0.341)	Data 0.000 (0.074)	Loss 0.845 (0.696)
Epoch: [48][140/200]	Time 0.258 (0.339)	Data 0.001 (0.073)	Loss 1.015 (0.715)
Epoch: [48][160/200]	Time 0.259 (0.339)	Data 0.001 (0.072)	Loss 0.999 (0.723)
Epoch: [48][180/200]	Time 0.258 (0.338)	Data 0.001 (0.072)	Loss 0.709 (0.731)
Epoch: [48][200/200]	Time 0.261 (0.338)	Data 0.001 (0.072)	Loss 0.700 (0.742)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.095 (0.135)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.017266511917114
==> Statistics for epoch 49: 611 clusters
Epoch: [49][20/200]	Time 1.507 (0.373)	Data 1.207 (0.100)	Loss 0.688 (0.137)
Epoch: [49][40/200]	Time 0.258 (0.351)	Data 0.001 (0.082)	Loss 0.753 (0.481)
Epoch: [49][60/200]	Time 0.256 (0.344)	Data 0.001 (0.076)	Loss 1.003 (0.599)
Epoch: [49][80/200]	Time 0.257 (0.339)	Data 0.000 (0.073)	Loss 0.724 (0.652)
Epoch: [49][100/200]	Time 0.260 (0.338)	Data 0.001 (0.072)	Loss 0.465 (0.689)
Epoch: [49][120/200]	Time 0.262 (0.336)	Data 0.001 (0.070)	Loss 0.632 (0.708)
Epoch: [49][140/200]	Time 0.256 (0.335)	Data 0.001 (0.069)	Loss 0.875 (0.721)
Epoch: [49][160/200]	Time 0.256 (0.334)	Data 0.001 (0.068)	Loss 0.708 (0.732)
Epoch: [49][180/200]	Time 0.262 (0.333)	Data 0.001 (0.068)	Loss 0.810 (0.742)
Epoch: [49][200/200]	Time 0.365 (0.332)	Data 0.001 (0.067)	Loss 0.787 (0.745)
Extract Features: [50/76]	Time 0.093 (0.130)	Data 0.000 (0.034)	
Mean AP: 92.1%

 * Finished epoch  49  model mAP: 92.1%  best: 92.1%

==> Test with the best model:
=> Loaded checkpoint 'log/msmt2market/resnet50_ibn_cion/model_best.pth.tar'
Extract Features: [50/76]	Time 0.093 (0.131)	Data 0.000 (0.032)	
Mean AP: 92.1%
CMC Scores:
  top-1          96.5%
  top-5          98.7%
  top-10         99.3%
Total running time:  1:23:28.792975
