==========
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='resnet101', pretrained_path='/home/ma-user/work/Projects/ReIDNet_Finetune/TransReID/log/bs_exp/ResNet101_MSMT17/64bs_lr0.0004_ep120_warm20_seed0/resnet101_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_cion', pooling_type='gem', use_hard=True)
==========
==> Load unlabeled dataset
=> Market1501 loaded
Dataset statistics:
  ----------------------------------------
  subset   | # ids | # images | # cameras
  ----------------------------------------
  train    |   751 |    12936 |         6
  query    |   750 |     3368 |         6
  gallery  |   751 |    15913 |         6
  ----------------------------------------
pooling_type: gem
optimizer: SGD
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.520)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 21.825780153274536
Clustering criterion: eps: 0.600
==> Statistics for epoch 0: 548 clusters
Epoch: [0][20/200]	Time 0.377 (0.885)	Data 0.001 (0.100)	Loss 2.575 (3.285)
Epoch: [0][40/200]	Time 0.377 (0.668)	Data 0.001 (0.083)	Loss 2.314 (3.029)
Epoch: [0][60/200]	Time 0.378 (0.596)	Data 0.001 (0.076)	Loss 2.169 (2.779)
Epoch: [0][80/200]	Time 0.376 (0.558)	Data 0.000 (0.073)	Loss 2.235 (2.615)
Epoch: [0][100/200]	Time 0.377 (0.537)	Data 0.000 (0.070)	Loss 1.843 (2.491)
Epoch: [0][120/200]	Time 1.757 (0.536)	Data 1.304 (0.081)	Loss 2.194 (2.395)
Epoch: [0][140/200]	Time 0.378 (0.526)	Data 0.001 (0.079)	Loss 1.972 (2.325)
Epoch: [0][160/200]	Time 0.379 (0.516)	Data 0.001 (0.077)	Loss 2.184 (2.262)
Epoch: [0][180/200]	Time 0.380 (0.509)	Data 0.000 (0.076)	Loss 2.152 (2.213)
Epoch: [0][200/200]	Time 0.376 (0.503)	Data 0.000 (0.075)	Loss 1.633 (2.168)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.171)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.708393335342407
==> Statistics for epoch 1: 604 clusters
Epoch: [1][20/200]	Time 0.384 (0.502)	Data 0.001 (0.111)	Loss 1.688 (0.426)
Epoch: [1][40/200]	Time 0.383 (0.480)	Data 0.001 (0.089)	Loss 1.865 (1.016)
Epoch: [1][60/200]	Time 0.380 (0.473)	Data 0.001 (0.083)	Loss 1.694 (1.235)
Epoch: [1][80/200]	Time 0.377 (0.469)	Data 0.001 (0.078)	Loss 1.668 (1.338)
Epoch: [1][100/200]	Time 0.381 (0.466)	Data 0.001 (0.076)	Loss 1.473 (1.416)
Epoch: [1][120/200]	Time 0.377 (0.465)	Data 0.000 (0.074)	Loss 1.516 (1.455)
Epoch: [1][140/200]	Time 0.378 (0.464)	Data 0.000 (0.073)	Loss 1.740 (1.486)
Epoch: [1][160/200]	Time 0.447 (0.463)	Data 0.000 (0.072)	Loss 1.559 (1.503)
Epoch: [1][180/200]	Time 0.375 (0.462)	Data 0.000 (0.071)	Loss 1.699 (1.509)
Epoch: [1][200/200]	Time 0.388 (0.468)	Data 0.001 (0.077)	Loss 1.460 (1.517)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.174)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.73551893234253
==> Statistics for epoch 2: 607 clusters
Epoch: [2][20/200]	Time 0.385 (0.505)	Data 0.001 (0.112)	Loss 1.488 (0.407)
Epoch: [2][40/200]	Time 0.384 (0.488)	Data 0.001 (0.096)	Loss 1.353 (0.959)
Epoch: [2][60/200]	Time 0.379 (0.480)	Data 0.002 (0.088)	Loss 1.594 (1.134)
Epoch: [2][80/200]	Time 0.381 (0.475)	Data 0.001 (0.084)	Loss 1.563 (1.223)
Epoch: [2][100/200]	Time 0.382 (0.473)	Data 0.001 (0.081)	Loss 1.331 (1.269)
Epoch: [2][120/200]	Time 0.385 (0.470)	Data 0.000 (0.078)	Loss 1.365 (1.304)
Epoch: [2][140/200]	Time 0.379 (0.468)	Data 0.000 (0.077)	Loss 1.222 (1.325)
Epoch: [2][160/200]	Time 0.375 (0.467)	Data 0.000 (0.075)	Loss 1.179 (1.345)
Epoch: [2][180/200]	Time 0.378 (0.466)	Data 0.000 (0.074)	Loss 1.376 (1.359)
Epoch: [2][200/200]	Time 0.393 (0.474)	Data 0.001 (0.082)	Loss 1.134 (1.357)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.173)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.197632551193237
==> Statistics for epoch 3: 610 clusters
Epoch: [3][20/200]	Time 1.953 (0.508)	Data 1.563 (0.117)	Loss 1.495 (0.288)
Epoch: [3][40/200]	Time 0.390 (0.481)	Data 0.001 (0.090)	Loss 1.561 (0.849)
Epoch: [3][60/200]	Time 0.383 (0.472)	Data 0.001 (0.082)	Loss 1.271 (1.019)
Epoch: [3][80/200]	Time 0.379 (0.472)	Data 0.001 (0.081)	Loss 1.326 (1.117)
Epoch: [3][100/200]	Time 0.377 (0.470)	Data 0.001 (0.078)	Loss 1.348 (1.167)
Epoch: [3][120/200]	Time 0.383 (0.470)	Data 0.001 (0.077)	Loss 1.299 (1.195)
Epoch: [3][140/200]	Time 0.380 (0.469)	Data 0.001 (0.076)	Loss 1.290 (1.219)
Epoch: [3][160/200]	Time 0.381 (0.470)	Data 0.001 (0.076)	Loss 1.673 (1.234)
Epoch: [3][180/200]	Time 0.381 (0.470)	Data 0.001 (0.077)	Loss 1.361 (1.239)
Epoch: [3][200/200]	Time 0.386 (0.470)	Data 0.002 (0.077)	Loss 1.138 (1.250)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.172)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.00823950767517
==> Statistics for epoch 4: 617 clusters
Epoch: [4][20/200]	Time 1.764 (0.495)	Data 1.349 (0.104)	Loss 1.131 (0.247)
Epoch: [4][40/200]	Time 0.389 (0.476)	Data 0.001 (0.086)	Loss 1.025 (0.784)
Epoch: [4][60/200]	Time 0.384 (0.471)	Data 0.001 (0.081)	Loss 1.241 (0.960)
Epoch: [4][80/200]	Time 0.379 (0.469)	Data 0.001 (0.077)	Loss 1.006 (1.037)
Epoch: [4][100/200]	Time 0.381 (0.467)	Data 0.001 (0.075)	Loss 1.149 (1.095)
Epoch: [4][120/200]	Time 0.383 (0.465)	Data 0.001 (0.074)	Loss 1.230 (1.138)
Epoch: [4][140/200]	Time 0.381 (0.464)	Data 0.001 (0.073)	Loss 1.569 (1.162)
Epoch: [4][160/200]	Time 0.384 (0.464)	Data 0.001 (0.073)	Loss 1.309 (1.172)
Epoch: [4][180/200]	Time 0.394 (0.463)	Data 0.001 (0.072)	Loss 1.555 (1.186)
Epoch: [4][200/200]	Time 0.378 (0.463)	Data 0.001 (0.072)	Loss 1.659 (1.197)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.141 (0.171)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.030217170715332
==> Statistics for epoch 5: 612 clusters
Epoch: [5][20/200]	Time 1.721 (0.492)	Data 1.318 (0.108)	Loss 1.518 (0.243)
Epoch: [5][40/200]	Time 0.386 (0.473)	Data 0.001 (0.087)	Loss 1.207 (0.707)
Epoch: [5][60/200]	Time 0.377 (0.468)	Data 0.001 (0.080)	Loss 1.261 (0.882)
Epoch: [5][80/200]	Time 0.380 (0.465)	Data 0.001 (0.076)	Loss 1.172 (0.970)
Epoch: [5][100/200]	Time 0.388 (0.463)	Data 0.001 (0.074)	Loss 1.510 (1.029)
Epoch: [5][120/200]	Time 0.381 (0.462)	Data 0.001 (0.072)	Loss 1.372 (1.056)
Epoch: [5][140/200]	Time 0.381 (0.461)	Data 0.001 (0.071)	Loss 1.223 (1.081)
Epoch: [5][160/200]	Time 0.378 (0.462)	Data 0.001 (0.072)	Loss 1.193 (1.097)
Epoch: [5][180/200]	Time 0.380 (0.461)	Data 0.001 (0.070)	Loss 1.135 (1.114)
Epoch: [5][200/200]	Time 0.384 (0.460)	Data 0.001 (0.070)	Loss 1.382 (1.128)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.169)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.135501861572266
==> Statistics for epoch 6: 617 clusters
Epoch: [6][20/200]	Time 1.968 (0.511)	Data 1.554 (0.117)	Loss 0.954 (0.209)
Epoch: [6][40/200]	Time 0.377 (0.481)	Data 0.001 (0.091)	Loss 1.367 (0.706)
Epoch: [6][60/200]	Time 0.380 (0.473)	Data 0.001 (0.084)	Loss 0.953 (0.847)
Epoch: [6][80/200]	Time 0.378 (0.469)	Data 0.001 (0.079)	Loss 1.440 (0.936)
Epoch: [6][100/200]	Time 0.379 (0.465)	Data 0.001 (0.076)	Loss 1.182 (0.990)
Epoch: [6][120/200]	Time 0.375 (0.465)	Data 0.001 (0.076)	Loss 1.205 (1.014)
Epoch: [6][140/200]	Time 0.378 (0.463)	Data 0.001 (0.074)	Loss 1.027 (1.037)
Epoch: [6][160/200]	Time 0.382 (0.464)	Data 0.001 (0.074)	Loss 1.673 (1.060)
Epoch: [6][180/200]	Time 0.383 (0.463)	Data 0.001 (0.073)	Loss 1.010 (1.064)
Epoch: [6][200/200]	Time 0.383 (0.464)	Data 0.001 (0.073)	Loss 1.099 (1.069)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.179)	Data 0.000 (0.031)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.770620346069336
==> Statistics for epoch 7: 618 clusters
Epoch: [7][20/200]	Time 1.841 (0.505)	Data 1.419 (0.117)	Loss 0.990 (0.208)
Epoch: [7][40/200]	Time 0.392 (0.490)	Data 0.001 (0.101)	Loss 0.735 (0.649)
Epoch: [7][60/200]	Time 0.385 (0.481)	Data 0.001 (0.091)	Loss 1.427 (0.817)
Epoch: [7][80/200]	Time 0.383 (0.476)	Data 0.001 (0.087)	Loss 1.125 (0.893)
Epoch: [7][100/200]	Time 0.379 (0.475)	Data 0.001 (0.085)	Loss 1.360 (0.942)
Epoch: [7][120/200]	Time 0.378 (0.474)	Data 0.001 (0.083)	Loss 1.047 (0.966)
Epoch: [7][140/200]	Time 0.379 (0.472)	Data 0.001 (0.081)	Loss 1.020 (0.986)
Epoch: [7][160/200]	Time 0.379 (0.471)	Data 0.001 (0.080)	Loss 1.432 (1.006)
Epoch: [7][180/200]	Time 0.379 (0.471)	Data 0.001 (0.080)	Loss 1.130 (1.007)
Epoch: [7][200/200]	Time 0.380 (0.471)	Data 0.001 (0.080)	Loss 1.176 (1.017)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.144 (0.176)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.749812841415405
==> Statistics for epoch 8: 614 clusters
Epoch: [8][20/200]	Time 2.047 (0.508)	Data 1.457 (0.108)	Loss 0.836 (0.185)
Epoch: [8][40/200]	Time 0.387 (0.484)	Data 0.001 (0.090)	Loss 0.945 (0.616)
Epoch: [8][60/200]	Time 0.387 (0.479)	Data 0.001 (0.086)	Loss 0.862 (0.767)
Epoch: [8][80/200]	Time 0.380 (0.474)	Data 0.001 (0.082)	Loss 1.078 (0.851)
Epoch: [8][100/200]	Time 0.382 (0.472)	Data 0.001 (0.080)	Loss 1.406 (0.899)
Epoch: [8][120/200]	Time 0.381 (0.471)	Data 0.001 (0.079)	Loss 1.002 (0.928)
Epoch: [8][140/200]	Time 0.383 (0.471)	Data 0.001 (0.078)	Loss 0.835 (0.957)
Epoch: [8][160/200]	Time 0.381 (0.471)	Data 0.001 (0.078)	Loss 0.683 (0.963)
Epoch: [8][180/200]	Time 0.380 (0.470)	Data 0.001 (0.078)	Loss 0.844 (0.966)
Epoch: [8][200/200]	Time 0.380 (0.468)	Data 0.001 (0.077)	Loss 0.879 (0.976)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.140 (0.175)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.856992721557617
==> Statistics for epoch 9: 614 clusters
Epoch: [9][20/200]	Time 1.699 (0.487)	Data 1.274 (0.101)	Loss 0.986 (0.186)
Epoch: [9][40/200]	Time 0.381 (0.473)	Data 0.001 (0.082)	Loss 0.762 (0.587)
Epoch: [9][60/200]	Time 0.377 (0.470)	Data 0.001 (0.082)	Loss 1.022 (0.758)
Epoch: [9][80/200]	Time 0.382 (0.469)	Data 0.001 (0.080)	Loss 0.566 (0.834)
Epoch: [9][100/200]	Time 0.380 (0.467)	Data 0.001 (0.077)	Loss 1.119 (0.891)
Epoch: [9][120/200]	Time 0.385 (0.465)	Data 0.001 (0.075)	Loss 0.931 (0.920)
Epoch: [9][140/200]	Time 0.385 (0.464)	Data 0.001 (0.074)	Loss 0.813 (0.940)
Epoch: [9][160/200]	Time 0.387 (0.463)	Data 0.001 (0.073)	Loss 0.772 (0.945)
Epoch: [9][180/200]	Time 0.479 (0.465)	Data 0.001 (0.073)	Loss 0.788 (0.959)
Epoch: [9][200/200]	Time 0.479 (0.465)	Data 0.001 (0.073)	Loss 1.146 (0.959)
Extract Features: [50/76]	Time 0.139 (0.168)	Data 0.000 (0.023)	
Mean AP: 92.2%

 * Finished epoch   9  model mAP: 92.2%  best: 92.2% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.140 (0.171)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.610499143600464
==> Statistics for epoch 10: 619 clusters
Epoch: [10][20/200]	Time 1.876 (0.532)	Data 1.440 (0.108)	Loss 0.962 (0.197)
Epoch: [10][40/200]	Time 0.387 (0.498)	Data 0.001 (0.090)	Loss 0.924 (0.589)
Epoch: [10][60/200]	Time 0.385 (0.489)	Data 0.001 (0.087)	Loss 0.860 (0.730)
Epoch: [10][80/200]	Time 0.379 (0.485)	Data 0.002 (0.085)	Loss 0.874 (0.799)
Epoch: [10][100/200]	Time 0.383 (0.482)	Data 0.001 (0.083)	Loss 1.345 (0.838)
Epoch: [10][120/200]	Time 0.386 (0.478)	Data 0.002 (0.081)	Loss 1.082 (0.862)
Epoch: [10][140/200]	Time 0.383 (0.475)	Data 0.001 (0.079)	Loss 0.946 (0.891)
Epoch: [10][160/200]	Time 0.382 (0.474)	Data 0.001 (0.078)	Loss 0.967 (0.908)
Epoch: [10][180/200]	Time 0.383 (0.473)	Data 0.001 (0.077)	Loss 0.919 (0.915)
Epoch: [10][200/200]	Time 0.380 (0.473)	Data 0.001 (0.077)	Loss 1.117 (0.936)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.171)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.994219064712524
==> Statistics for epoch 11: 625 clusters
Epoch: [11][20/200]	Time 1.867 (0.500)	Data 1.441 (0.109)	Loss 1.130 (0.198)
Epoch: [11][40/200]	Time 0.389 (0.483)	Data 0.001 (0.090)	Loss 0.938 (0.565)
Epoch: [11][60/200]	Time 0.384 (0.477)	Data 0.001 (0.085)	Loss 0.862 (0.696)
Epoch: [11][80/200]	Time 0.379 (0.476)	Data 0.001 (0.086)	Loss 1.078 (0.786)
Epoch: [11][100/200]	Time 0.380 (0.473)	Data 0.001 (0.083)	Loss 1.279 (0.838)
Epoch: [11][120/200]	Time 0.380 (0.470)	Data 0.001 (0.080)	Loss 0.949 (0.859)
Epoch: [11][140/200]	Time 0.381 (0.470)	Data 0.001 (0.079)	Loss 1.082 (0.879)
Epoch: [11][160/200]	Time 0.378 (0.469)	Data 0.001 (0.078)	Loss 0.709 (0.890)
Epoch: [11][180/200]	Time 0.381 (0.468)	Data 0.001 (0.077)	Loss 0.931 (0.900)
Epoch: [11][200/200]	Time 0.507 (0.468)	Data 0.001 (0.077)	Loss 0.863 (0.907)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.171)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.917449474334717
==> Statistics for epoch 12: 621 clusters
Epoch: [12][20/200]	Time 1.884 (0.501)	Data 1.376 (0.104)	Loss 0.820 (0.157)
Epoch: [12][40/200]	Time 0.488 (0.480)	Data 0.001 (0.087)	Loss 0.952 (0.519)
Epoch: [12][60/200]	Time 0.380 (0.470)	Data 0.001 (0.079)	Loss 0.834 (0.655)
Epoch: [12][80/200]	Time 0.376 (0.467)	Data 0.001 (0.076)	Loss 0.890 (0.725)
Epoch: [12][100/200]	Time 0.382 (0.465)	Data 0.001 (0.074)	Loss 0.929 (0.782)
Epoch: [12][120/200]	Time 0.382 (0.463)	Data 0.001 (0.072)	Loss 1.133 (0.803)
Epoch: [12][140/200]	Time 0.380 (0.461)	Data 0.001 (0.071)	Loss 0.684 (0.819)
Epoch: [12][160/200]	Time 0.386 (0.460)	Data 0.001 (0.069)	Loss 0.996 (0.842)
Epoch: [12][180/200]	Time 0.384 (0.460)	Data 0.001 (0.069)	Loss 0.798 (0.847)
Epoch: [12][200/200]	Time 0.379 (0.459)	Data 0.001 (0.068)	Loss 1.044 (0.861)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.169)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.56475520133972
==> Statistics for epoch 13: 623 clusters
Epoch: [13][20/200]	Time 1.891 (0.509)	Data 1.480 (0.114)	Loss 0.897 (0.172)
Epoch: [13][40/200]	Time 0.390 (0.488)	Data 0.001 (0.093)	Loss 0.827 (0.519)
Epoch: [13][60/200]	Time 0.377 (0.478)	Data 0.001 (0.086)	Loss 1.000 (0.640)
Epoch: [13][80/200]	Time 0.380 (0.475)	Data 0.001 (0.082)	Loss 0.798 (0.717)
Epoch: [13][100/200]	Time 0.382 (0.473)	Data 0.001 (0.080)	Loss 0.646 (0.766)
Epoch: [13][120/200]	Time 0.383 (0.471)	Data 0.001 (0.078)	Loss 0.993 (0.791)
Epoch: [13][140/200]	Time 0.386 (0.470)	Data 0.001 (0.078)	Loss 0.678 (0.810)
Epoch: [13][160/200]	Time 0.381 (0.470)	Data 0.001 (0.077)	Loss 0.748 (0.824)
Epoch: [13][180/200]	Time 0.383 (0.469)	Data 0.001 (0.076)	Loss 0.608 (0.831)
Epoch: [13][200/200]	Time 0.384 (0.468)	Data 0.001 (0.075)	Loss 0.891 (0.844)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.238 (0.174)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.651523113250732
==> Statistics for epoch 14: 619 clusters
Epoch: [14][20/200]	Time 1.728 (0.501)	Data 1.311 (0.103)	Loss 0.700 (0.150)
Epoch: [14][40/200]	Time 0.382 (0.477)	Data 0.001 (0.083)	Loss 0.796 (0.510)
Epoch: [14][60/200]	Time 0.381 (0.468)	Data 0.001 (0.076)	Loss 0.956 (0.645)
Epoch: [14][80/200]	Time 0.378 (0.466)	Data 0.001 (0.072)	Loss 0.666 (0.689)
Epoch: [14][100/200]	Time 0.379 (0.463)	Data 0.001 (0.071)	Loss 0.802 (0.728)
Epoch: [14][120/200]	Time 0.468 (0.462)	Data 0.001 (0.070)	Loss 0.868 (0.751)
Epoch: [14][140/200]	Time 0.383 (0.461)	Data 0.001 (0.069)	Loss 0.807 (0.767)
Epoch: [14][160/200]	Time 0.381 (0.460)	Data 0.001 (0.069)	Loss 0.785 (0.787)
Epoch: [14][180/200]	Time 0.381 (0.459)	Data 0.001 (0.068)	Loss 0.877 (0.795)
Epoch: [14][200/200]	Time 0.383 (0.459)	Data 0.001 (0.068)	Loss 0.767 (0.806)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.166)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.889429092407227
==> Statistics for epoch 15: 622 clusters
Epoch: [15][20/200]	Time 1.816 (0.503)	Data 1.388 (0.105)	Loss 0.905 (0.174)
Epoch: [15][40/200]	Time 0.381 (0.487)	Data 0.001 (0.092)	Loss 1.214 (0.533)
Epoch: [15][60/200]	Time 0.383 (0.481)	Data 0.001 (0.085)	Loss 0.990 (0.631)
Epoch: [15][80/200]	Time 0.381 (0.475)	Data 0.001 (0.080)	Loss 0.711 (0.693)
Epoch: [15][100/200]	Time 0.381 (0.473)	Data 0.001 (0.078)	Loss 0.640 (0.729)
Epoch: [15][120/200]	Time 0.501 (0.472)	Data 0.001 (0.078)	Loss 1.031 (0.751)
Epoch: [15][140/200]	Time 0.387 (0.471)	Data 0.001 (0.077)	Loss 1.015 (0.768)
Epoch: [15][160/200]	Time 0.385 (0.470)	Data 0.001 (0.076)	Loss 0.965 (0.778)
Epoch: [15][180/200]	Time 0.380 (0.470)	Data 0.001 (0.076)	Loss 0.894 (0.784)
Epoch: [15][200/200]	Time 0.383 (0.470)	Data 0.001 (0.076)	Loss 1.089 (0.789)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.173)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.07200598716736
==> Statistics for epoch 16: 624 clusters
Epoch: [16][20/200]	Time 1.895 (0.506)	Data 1.442 (0.110)	Loss 1.067 (0.179)
Epoch: [16][40/200]	Time 0.381 (0.487)	Data 0.001 (0.094)	Loss 1.344 (0.525)
Epoch: [16][60/200]	Time 0.382 (0.476)	Data 0.001 (0.083)	Loss 1.121 (0.644)
Epoch: [16][80/200]	Time 0.381 (0.471)	Data 0.001 (0.079)	Loss 1.031 (0.706)
Epoch: [16][100/200]	Time 0.379 (0.467)	Data 0.001 (0.076)	Loss 0.778 (0.731)
Epoch: [16][120/200]	Time 0.380 (0.464)	Data 0.001 (0.074)	Loss 0.979 (0.756)
Epoch: [16][140/200]	Time 0.380 (0.463)	Data 0.001 (0.073)	Loss 0.675 (0.766)
Epoch: [16][160/200]	Time 0.382 (0.462)	Data 0.000 (0.071)	Loss 0.920 (0.776)
Epoch: [16][180/200]	Time 0.386 (0.461)	Data 0.001 (0.070)	Loss 0.699 (0.782)
Epoch: [16][200/200]	Time 0.460 (0.460)	Data 0.001 (0.070)	Loss 0.733 (0.790)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.167)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.88948941230774
==> Statistics for epoch 17: 619 clusters
Epoch: [17][20/200]	Time 1.778 (0.496)	Data 1.267 (0.105)	Loss 0.771 (0.144)
Epoch: [17][40/200]	Time 0.393 (0.475)	Data 0.001 (0.087)	Loss 0.587 (0.458)
Epoch: [17][60/200]	Time 0.377 (0.467)	Data 0.001 (0.080)	Loss 1.071 (0.585)
Epoch: [17][80/200]	Time 0.378 (0.463)	Data 0.001 (0.076)	Loss 0.909 (0.652)
Epoch: [17][100/200]	Time 0.381 (0.461)	Data 0.001 (0.073)	Loss 0.785 (0.687)
Epoch: [17][120/200]	Time 0.379 (0.460)	Data 0.001 (0.071)	Loss 0.791 (0.712)
Epoch: [17][140/200]	Time 0.379 (0.458)	Data 0.001 (0.070)	Loss 0.566 (0.728)
Epoch: [17][160/200]	Time 0.379 (0.458)	Data 0.001 (0.069)	Loss 0.673 (0.742)
Epoch: [17][180/200]	Time 0.382 (0.457)	Data 0.001 (0.068)	Loss 0.846 (0.747)
Epoch: [17][200/200]	Time 0.378 (0.457)	Data 0.001 (0.067)	Loss 1.218 (0.763)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.169)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.64040446281433
==> Statistics for epoch 18: 619 clusters
Epoch: [18][20/200]	Time 1.782 (0.506)	Data 1.361 (0.109)	Loss 0.797 (0.147)
Epoch: [18][40/200]	Time 0.389 (0.477)	Data 0.001 (0.086)	Loss 1.152 (0.499)
Epoch: [18][60/200]	Time 0.384 (0.471)	Data 0.001 (0.078)	Loss 0.709 (0.608)
Epoch: [18][80/200]	Time 0.478 (0.467)	Data 0.001 (0.074)	Loss 1.017 (0.670)
Epoch: [18][100/200]	Time 0.506 (0.465)	Data 0.001 (0.072)	Loss 1.156 (0.706)
Epoch: [18][120/200]	Time 0.378 (0.463)	Data 0.001 (0.071)	Loss 0.588 (0.724)
Epoch: [18][140/200]	Time 0.384 (0.462)	Data 0.001 (0.070)	Loss 0.671 (0.733)
Epoch: [18][160/200]	Time 0.380 (0.462)	Data 0.001 (0.070)	Loss 0.745 (0.742)
Epoch: [18][180/200]	Time 0.478 (0.462)	Data 0.001 (0.069)	Loss 0.767 (0.746)
Epoch: [18][200/200]	Time 0.485 (0.462)	Data 0.001 (0.069)	Loss 0.882 (0.755)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.167)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.012338876724243
==> Statistics for epoch 19: 621 clusters
Epoch: [19][20/200]	Time 1.834 (0.508)	Data 1.437 (0.115)	Loss 0.989 (0.157)
Epoch: [19][40/200]	Time 0.514 (0.484)	Data 0.001 (0.091)	Loss 0.771 (0.458)
Epoch: [19][60/200]	Time 0.381 (0.476)	Data 0.001 (0.084)	Loss 0.688 (0.571)
Epoch: [19][80/200]	Time 0.378 (0.474)	Data 0.001 (0.083)	Loss 0.622 (0.625)
Epoch: [19][100/200]	Time 0.381 (0.471)	Data 0.001 (0.080)	Loss 0.885 (0.659)
Epoch: [19][120/200]	Time 0.381 (0.470)	Data 0.001 (0.078)	Loss 0.540 (0.673)
Epoch: [19][140/200]	Time 0.383 (0.469)	Data 0.001 (0.077)	Loss 0.644 (0.688)
Epoch: [19][160/200]	Time 0.393 (0.467)	Data 0.001 (0.075)	Loss 0.653 (0.704)
Epoch: [19][180/200]	Time 0.389 (0.467)	Data 0.001 (0.074)	Loss 0.828 (0.714)
Epoch: [19][200/200]	Time 0.382 (0.468)	Data 0.001 (0.075)	Loss 0.716 (0.720)
Extract Features: [50/76]	Time 0.137 (0.171)	Data 0.000 (0.025)	
Mean AP: 92.9%

 * Finished epoch  19  model mAP: 92.9%  best: 92.9% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.171)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.801862478256226
==> Statistics for epoch 20: 620 clusters
Epoch: [20][20/200]	Time 1.683 (0.495)	Data 1.258 (0.097)	Loss 0.576 (0.119)
Epoch: [20][40/200]	Time 0.393 (0.473)	Data 0.001 (0.079)	Loss 0.875 (0.416)
Epoch: [20][60/200]	Time 0.376 (0.466)	Data 0.001 (0.073)	Loss 0.575 (0.527)
Epoch: [20][80/200]	Time 0.376 (0.462)	Data 0.001 (0.070)	Loss 0.575 (0.596)
Epoch: [20][100/200]	Time 0.379 (0.461)	Data 0.001 (0.069)	Loss 0.797 (0.627)
Epoch: [20][120/200]	Time 0.386 (0.459)	Data 0.001 (0.068)	Loss 0.592 (0.656)
Epoch: [20][140/200]	Time 0.497 (0.459)	Data 0.001 (0.068)	Loss 0.630 (0.672)
Epoch: [20][160/200]	Time 0.379 (0.458)	Data 0.001 (0.067)	Loss 0.860 (0.683)
Epoch: [20][180/200]	Time 0.378 (0.457)	Data 0.001 (0.067)	Loss 0.813 (0.689)
Epoch: [20][200/200]	Time 0.379 (0.457)	Data 0.001 (0.066)	Loss 0.533 (0.692)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.165)	Data 0.000 (0.020)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.553555727005005
==> Statistics for epoch 21: 621 clusters
Epoch: [21][20/200]	Time 1.947 (0.510)	Data 1.500 (0.111)	Loss 0.811 (0.147)
Epoch: [21][40/200]	Time 0.381 (0.489)	Data 0.001 (0.094)	Loss 0.832 (0.450)
Epoch: [21][60/200]	Time 0.490 (0.484)	Data 0.001 (0.088)	Loss 1.175 (0.547)
Epoch: [21][80/200]	Time 0.377 (0.477)	Data 0.001 (0.082)	Loss 0.837 (0.601)
Epoch: [21][100/200]	Time 0.488 (0.474)	Data 0.001 (0.079)	Loss 0.581 (0.640)
Epoch: [21][120/200]	Time 0.379 (0.471)	Data 0.001 (0.077)	Loss 0.573 (0.663)
Epoch: [21][140/200]	Time 0.380 (0.470)	Data 0.001 (0.077)	Loss 0.809 (0.674)
Epoch: [21][160/200]	Time 0.380 (0.469)	Data 0.001 (0.076)	Loss 0.435 (0.680)
Epoch: [21][180/200]	Time 0.380 (0.467)	Data 0.001 (0.074)	Loss 0.824 (0.689)
Epoch: [21][200/200]	Time 0.382 (0.466)	Data 0.001 (0.073)	Loss 0.774 (0.695)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.170)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.995698928833008
==> Statistics for epoch 22: 621 clusters
Epoch: [22][20/200]	Time 1.730 (0.498)	Data 1.304 (0.105)	Loss 0.939 (0.149)
Epoch: [22][40/200]	Time 0.397 (0.482)	Data 0.001 (0.090)	Loss 0.869 (0.452)
Epoch: [22][60/200]	Time 0.380 (0.476)	Data 0.001 (0.083)	Loss 1.028 (0.562)
Epoch: [22][80/200]	Time 0.487 (0.472)	Data 0.001 (0.079)	Loss 0.694 (0.615)
Epoch: [22][100/200]	Time 0.502 (0.469)	Data 0.001 (0.076)	Loss 0.628 (0.640)
Epoch: [22][120/200]	Time 0.380 (0.469)	Data 0.001 (0.076)	Loss 0.784 (0.660)
Epoch: [22][140/200]	Time 0.382 (0.469)	Data 0.001 (0.075)	Loss 0.688 (0.670)
Epoch: [22][160/200]	Time 0.382 (0.467)	Data 0.001 (0.074)	Loss 0.654 (0.677)
Epoch: [22][180/200]	Time 0.381 (0.467)	Data 0.001 (0.074)	Loss 1.122 (0.685)
Epoch: [22][200/200]	Time 0.380 (0.466)	Data 0.001 (0.073)	Loss 0.501 (0.692)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.171)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.728883981704712
==> Statistics for epoch 23: 622 clusters
Epoch: [23][20/200]	Time 1.885 (0.512)	Data 1.449 (0.117)	Loss 0.830 (0.137)
Epoch: [23][40/200]	Time 0.389 (0.487)	Data 0.001 (0.094)	Loss 0.829 (0.424)
Epoch: [23][60/200]	Time 0.380 (0.478)	Data 0.001 (0.085)	Loss 1.132 (0.543)
Epoch: [23][80/200]	Time 0.380 (0.473)	Data 0.001 (0.081)	Loss 0.652 (0.579)
Epoch: [23][100/200]	Time 0.382 (0.470)	Data 0.001 (0.079)	Loss 0.624 (0.606)
Epoch: [23][120/200]	Time 0.379 (0.467)	Data 0.001 (0.076)	Loss 0.899 (0.634)
Epoch: [23][140/200]	Time 0.381 (0.465)	Data 0.001 (0.074)	Loss 1.303 (0.662)
Epoch: [23][160/200]	Time 0.382 (0.465)	Data 0.001 (0.073)	Loss 0.660 (0.672)
Epoch: [23][180/200]	Time 0.381 (0.464)	Data 0.001 (0.073)	Loss 0.813 (0.689)
Epoch: [23][200/200]	Time 0.480 (0.465)	Data 0.001 (0.073)	Loss 0.694 (0.691)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.170)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.11401653289795
==> Statistics for epoch 24: 620 clusters
Epoch: [24][20/200]	Time 1.769 (0.503)	Data 1.355 (0.109)	Loss 0.739 (0.130)
Epoch: [24][40/200]	Time 0.390 (0.485)	Data 0.000 (0.091)	Loss 0.687 (0.412)
Epoch: [24][60/200]	Time 0.380 (0.477)	Data 0.000 (0.083)	Loss 0.660 (0.527)
Epoch: [24][80/200]	Time 0.385 (0.475)	Data 0.001 (0.082)	Loss 0.885 (0.567)
Epoch: [24][100/200]	Time 0.379 (0.472)	Data 0.000 (0.079)	Loss 0.908 (0.605)
Epoch: [24][120/200]	Time 0.478 (0.470)	Data 0.000 (0.076)	Loss 0.714 (0.629)
Epoch: [24][140/200]	Time 0.378 (0.468)	Data 0.000 (0.075)	Loss 0.658 (0.646)
Epoch: [24][160/200]	Time 0.388 (0.466)	Data 0.000 (0.073)	Loss 0.579 (0.661)
Epoch: [24][180/200]	Time 0.382 (0.464)	Data 0.000 (0.072)	Loss 0.705 (0.673)
Epoch: [24][200/200]	Time 0.391 (0.465)	Data 0.001 (0.072)	Loss 0.690 (0.682)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.168)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.502442121505737
==> Statistics for epoch 25: 622 clusters
Epoch: [25][20/200]	Time 1.648 (0.491)	Data 1.224 (0.099)	Loss 0.634 (0.123)
Epoch: [25][40/200]	Time 0.388 (0.474)	Data 0.001 (0.082)	Loss 0.839 (0.413)
Epoch: [25][60/200]	Time 0.379 (0.469)	Data 0.001 (0.077)	Loss 0.756 (0.515)
Epoch: [25][80/200]	Time 0.381 (0.470)	Data 0.001 (0.077)	Loss 0.618 (0.568)
Epoch: [25][100/200]	Time 0.379 (0.468)	Data 0.001 (0.076)	Loss 0.688 (0.602)
Epoch: [25][120/200]	Time 0.382 (0.468)	Data 0.001 (0.076)	Loss 0.925 (0.629)
Epoch: [25][140/200]	Time 0.381 (0.467)	Data 0.001 (0.075)	Loss 1.014 (0.648)
Epoch: [25][160/200]	Time 0.479 (0.467)	Data 0.001 (0.075)	Loss 0.682 (0.662)
Epoch: [25][180/200]	Time 0.380 (0.467)	Data 0.001 (0.074)	Loss 0.499 (0.666)
Epoch: [25][200/200]	Time 0.459 (0.468)	Data 0.001 (0.075)	Loss 0.473 (0.671)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.171)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.345696449279785
==> Statistics for epoch 26: 620 clusters
Epoch: [26][20/200]	Time 1.687 (0.495)	Data 1.275 (0.100)	Loss 0.646 (0.124)
Epoch: [26][40/200]	Time 0.387 (0.475)	Data 0.001 (0.081)	Loss 0.749 (0.425)
Epoch: [26][60/200]	Time 0.377 (0.466)	Data 0.001 (0.074)	Loss 0.651 (0.508)
Epoch: [26][80/200]	Time 0.379 (0.463)	Data 0.001 (0.071)	Loss 0.664 (0.555)
Epoch: [26][100/200]	Time 0.389 (0.462)	Data 0.001 (0.070)	Loss 0.661 (0.593)
Epoch: [26][120/200]	Time 0.381 (0.460)	Data 0.001 (0.068)	Loss 0.702 (0.620)
Epoch: [26][140/200]	Time 0.384 (0.460)	Data 0.001 (0.068)	Loss 0.825 (0.638)
Epoch: [26][160/200]	Time 0.476 (0.460)	Data 0.001 (0.068)	Loss 1.124 (0.659)
Epoch: [26][180/200]	Time 0.380 (0.459)	Data 0.001 (0.068)	Loss 1.024 (0.667)
Epoch: [26][200/200]	Time 0.377 (0.459)	Data 0.001 (0.067)	Loss 0.594 (0.675)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.166)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.4013991355896
==> Statistics for epoch 27: 620 clusters
Epoch: [27][20/200]	Time 1.723 (0.503)	Data 1.293 (0.107)	Loss 0.535 (0.117)
Epoch: [27][40/200]	Time 0.381 (0.478)	Data 0.001 (0.083)	Loss 0.752 (0.412)
Epoch: [27][60/200]	Time 0.386 (0.470)	Data 0.001 (0.077)	Loss 1.053 (0.535)
Epoch: [27][80/200]	Time 0.377 (0.466)	Data 0.001 (0.075)	Loss 0.722 (0.595)
Epoch: [27][100/200]	Time 0.383 (0.464)	Data 0.001 (0.073)	Loss 0.615 (0.614)
Epoch: [27][120/200]	Time 0.383 (0.463)	Data 0.001 (0.073)	Loss 0.891 (0.629)
Epoch: [27][140/200]	Time 0.382 (0.462)	Data 0.001 (0.072)	Loss 0.681 (0.643)
Epoch: [27][160/200]	Time 0.377 (0.461)	Data 0.001 (0.071)	Loss 0.706 (0.655)
Epoch: [27][180/200]	Time 0.385 (0.461)	Data 0.001 (0.071)	Loss 0.747 (0.673)
Epoch: [27][200/200]	Time 0.375 (0.460)	Data 0.001 (0.070)	Loss 0.504 (0.675)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.168)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.914094924926758
==> Statistics for epoch 28: 620 clusters
Epoch: [28][20/200]	Time 1.681 (0.497)	Data 1.229 (0.098)	Loss 1.142 (0.157)
Epoch: [28][40/200]	Time 0.387 (0.475)	Data 0.001 (0.079)	Loss 0.420 (0.422)
Epoch: [28][60/200]	Time 0.385 (0.469)	Data 0.001 (0.074)	Loss 0.697 (0.520)
Epoch: [28][80/200]	Time 0.379 (0.468)	Data 0.001 (0.073)	Loss 0.698 (0.578)
Epoch: [28][100/200]	Time 0.385 (0.466)	Data 0.001 (0.071)	Loss 0.861 (0.598)
Epoch: [28][120/200]	Time 0.387 (0.464)	Data 0.001 (0.069)	Loss 0.781 (0.614)
Epoch: [28][140/200]	Time 0.377 (0.462)	Data 0.001 (0.069)	Loss 0.730 (0.626)
Epoch: [28][160/200]	Time 0.486 (0.464)	Data 0.001 (0.070)	Loss 0.612 (0.634)
Epoch: [28][180/200]	Time 0.378 (0.464)	Data 0.001 (0.070)	Loss 0.790 (0.645)
Epoch: [28][200/200]	Time 0.377 (0.463)	Data 0.001 (0.070)	Loss 1.026 (0.656)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.136 (0.170)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.650756359100342
==> Statistics for epoch 29: 619 clusters
Epoch: [29][20/200]	Time 1.775 (0.500)	Data 1.381 (0.107)	Loss 0.865 (0.136)
Epoch: [29][40/200]	Time 0.378 (0.486)	Data 0.001 (0.091)	Loss 0.628 (0.410)
Epoch: [29][60/200]	Time 0.380 (0.477)	Data 0.001 (0.082)	Loss 0.661 (0.508)
Epoch: [29][80/200]	Time 0.381 (0.471)	Data 0.001 (0.077)	Loss 0.737 (0.550)
Epoch: [29][100/200]	Time 0.396 (0.470)	Data 0.001 (0.076)	Loss 0.519 (0.582)
Epoch: [29][120/200]	Time 0.482 (0.468)	Data 0.001 (0.074)	Loss 0.706 (0.602)
Epoch: [29][140/200]	Time 0.400 (0.467)	Data 0.001 (0.073)	Loss 0.582 (0.624)
Epoch: [29][160/200]	Time 0.388 (0.467)	Data 0.001 (0.073)	Loss 0.817 (0.635)
Epoch: [29][180/200]	Time 0.381 (0.467)	Data 0.001 (0.073)	Loss 0.578 (0.640)
Epoch: [29][200/200]	Time 0.382 (0.467)	Data 0.001 (0.073)	Loss 0.640 (0.650)
Extract Features: [50/76]	Time 0.139 (0.176)	Data 0.000 (0.028)	
Mean AP: 93.0%

 * Finished epoch  29  model mAP: 93.0%  best: 93.0% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.168)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.767178297042847
==> Statistics for epoch 30: 620 clusters
Epoch: [30][20/200]	Time 1.902 (0.494)	Data 1.479 (0.110)	Loss 0.612 (0.115)
Epoch: [30][40/200]	Time 0.391 (0.479)	Data 0.001 (0.092)	Loss 0.710 (0.394)
Epoch: [30][60/200]	Time 0.387 (0.473)	Data 0.001 (0.086)	Loss 0.624 (0.498)
Epoch: [30][80/200]	Time 0.381 (0.472)	Data 0.001 (0.083)	Loss 0.741 (0.567)
Epoch: [30][100/200]	Time 0.382 (0.471)	Data 0.001 (0.081)	Loss 0.868 (0.613)
Epoch: [30][120/200]	Time 0.385 (0.470)	Data 0.001 (0.080)	Loss 0.730 (0.636)
Epoch: [30][140/200]	Time 0.385 (0.470)	Data 0.001 (0.079)	Loss 0.694 (0.645)
Epoch: [30][160/200]	Time 0.383 (0.469)	Data 0.001 (0.078)	Loss 0.959 (0.654)
Epoch: [30][180/200]	Time 0.382 (0.469)	Data 0.001 (0.078)	Loss 0.979 (0.661)
Epoch: [30][200/200]	Time 0.381 (0.469)	Data 0.001 (0.078)	Loss 0.713 (0.671)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.173)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.728172540664673
==> Statistics for epoch 31: 620 clusters
Epoch: [31][20/200]	Time 1.949 (0.518)	Data 1.533 (0.120)	Loss 0.617 (0.126)
Epoch: [31][40/200]	Time 0.386 (0.490)	Data 0.001 (0.094)	Loss 0.713 (0.400)
Epoch: [31][60/200]	Time 0.384 (0.481)	Data 0.001 (0.086)	Loss 0.699 (0.507)
Epoch: [31][80/200]	Time 0.387 (0.478)	Data 0.001 (0.083)	Loss 0.903 (0.557)
Epoch: [31][100/200]	Time 0.381 (0.476)	Data 0.001 (0.081)	Loss 0.710 (0.593)
Epoch: [31][120/200]	Time 0.382 (0.475)	Data 0.001 (0.079)	Loss 0.552 (0.609)
Epoch: [31][140/200]	Time 0.381 (0.473)	Data 0.001 (0.079)	Loss 0.624 (0.629)
Epoch: [31][160/200]	Time 0.384 (0.473)	Data 0.001 (0.078)	Loss 0.624 (0.632)
Epoch: [31][180/200]	Time 0.382 (0.472)	Data 0.001 (0.078)	Loss 0.955 (0.642)
Epoch: [31][200/200]	Time 0.490 (0.472)	Data 0.001 (0.078)	Loss 0.651 (0.651)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.172)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.445282459259033
==> Statistics for epoch 32: 621 clusters
Epoch: [32][20/200]	Time 1.972 (0.511)	Data 1.566 (0.113)	Loss 0.931 (0.129)
Epoch: [32][40/200]	Time 0.544 (0.489)	Data 0.001 (0.092)	Loss 0.628 (0.376)
Epoch: [32][60/200]	Time 0.384 (0.478)	Data 0.001 (0.083)	Loss 0.745 (0.488)
Epoch: [32][80/200]	Time 0.380 (0.476)	Data 0.001 (0.082)	Loss 0.494 (0.550)
Epoch: [32][100/200]	Time 0.476 (0.475)	Data 0.001 (0.081)	Loss 0.865 (0.575)
Epoch: [32][120/200]	Time 0.483 (0.473)	Data 0.001 (0.080)	Loss 0.589 (0.594)
Epoch: [32][140/200]	Time 0.381 (0.473)	Data 0.001 (0.080)	Loss 0.880 (0.620)
Epoch: [32][160/200]	Time 0.380 (0.472)	Data 0.001 (0.079)	Loss 0.753 (0.633)
Epoch: [32][180/200]	Time 0.383 (0.471)	Data 0.001 (0.078)	Loss 0.697 (0.645)
Epoch: [32][200/200]	Time 0.385 (0.471)	Data 0.001 (0.078)	Loss 1.059 (0.654)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.210 (0.176)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.656575918197632
==> Statistics for epoch 33: 619 clusters
Epoch: [33][20/200]	Time 1.922 (0.510)	Data 1.494 (0.115)	Loss 0.763 (0.117)
Epoch: [33][40/200]	Time 0.384 (0.487)	Data 0.001 (0.094)	Loss 0.659 (0.398)
Epoch: [33][60/200]	Time 0.384 (0.480)	Data 0.001 (0.087)	Loss 0.722 (0.497)
Epoch: [33][80/200]	Time 0.378 (0.476)	Data 0.001 (0.083)	Loss 0.997 (0.543)
Epoch: [33][100/200]	Time 0.387 (0.474)	Data 0.001 (0.082)	Loss 0.781 (0.575)
Epoch: [33][120/200]	Time 0.381 (0.473)	Data 0.001 (0.080)	Loss 0.724 (0.604)
Epoch: [33][140/200]	Time 0.381 (0.472)	Data 0.001 (0.080)	Loss 0.653 (0.622)
Epoch: [33][160/200]	Time 0.379 (0.471)	Data 0.001 (0.079)	Loss 0.749 (0.629)
Epoch: [33][180/200]	Time 0.381 (0.471)	Data 0.001 (0.079)	Loss 0.875 (0.637)
Epoch: [33][200/200]	Time 0.380 (0.471)	Data 0.001 (0.079)	Loss 0.380 (0.646)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.171)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.41841149330139
==> Statistics for epoch 34: 620 clusters
Epoch: [34][20/200]	Time 1.699 (0.498)	Data 1.282 (0.106)	Loss 0.388 (0.113)
Epoch: [34][40/200]	Time 0.393 (0.478)	Data 0.001 (0.088)	Loss 0.619 (0.399)
Epoch: [34][60/200]	Time 0.382 (0.471)	Data 0.001 (0.081)	Loss 0.835 (0.509)
Epoch: [34][80/200]	Time 0.378 (0.466)	Data 0.001 (0.076)	Loss 0.633 (0.573)
Epoch: [34][100/200]	Time 0.379 (0.464)	Data 0.001 (0.073)	Loss 0.547 (0.606)
Epoch: [34][120/200]	Time 0.384 (0.464)	Data 0.001 (0.074)	Loss 0.628 (0.624)
Epoch: [34][140/200]	Time 0.380 (0.463)	Data 0.001 (0.073)	Loss 0.735 (0.635)
Epoch: [34][160/200]	Time 0.381 (0.463)	Data 0.001 (0.072)	Loss 0.715 (0.647)
Epoch: [34][180/200]	Time 0.382 (0.463)	Data 0.001 (0.072)	Loss 0.671 (0.654)
Epoch: [34][200/200]	Time 0.383 (0.461)	Data 0.001 (0.071)	Loss 0.881 (0.661)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.168)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.392366409301758
==> Statistics for epoch 35: 618 clusters
Epoch: [35][20/200]	Time 1.678 (0.497)	Data 1.274 (0.105)	Loss 0.913 (0.130)
Epoch: [35][40/200]	Time 0.379 (0.478)	Data 0.001 (0.084)	Loss 0.664 (0.370)
Epoch: [35][60/200]	Time 0.384 (0.472)	Data 0.001 (0.080)	Loss 0.726 (0.487)
Epoch: [35][80/200]	Time 0.382 (0.466)	Data 0.001 (0.075)	Loss 0.680 (0.549)
Epoch: [35][100/200]	Time 0.381 (0.465)	Data 0.001 (0.074)	Loss 0.749 (0.581)
Epoch: [35][120/200]	Time 0.382 (0.466)	Data 0.001 (0.075)	Loss 0.839 (0.597)
Epoch: [35][140/200]	Time 0.492 (0.467)	Data 0.001 (0.074)	Loss 0.543 (0.612)
Epoch: [35][160/200]	Time 0.378 (0.466)	Data 0.001 (0.074)	Loss 0.621 (0.619)
Epoch: [35][180/200]	Time 0.380 (0.465)	Data 0.001 (0.073)	Loss 0.745 (0.628)
Epoch: [35][200/200]	Time 0.382 (0.464)	Data 0.001 (0.072)	Loss 0.903 (0.634)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.167)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.884886503219604
==> Statistics for epoch 36: 622 clusters
Epoch: [36][20/200]	Time 1.871 (0.514)	Data 1.343 (0.113)	Loss 0.500 (0.110)
Epoch: [36][40/200]	Time 0.396 (0.486)	Data 0.001 (0.091)	Loss 0.746 (0.387)
Epoch: [36][60/200]	Time 0.382 (0.480)	Data 0.001 (0.085)	Loss 0.584 (0.491)
Epoch: [36][80/200]	Time 0.382 (0.475)	Data 0.001 (0.080)	Loss 0.760 (0.544)
Epoch: [36][100/200]	Time 0.385 (0.472)	Data 0.001 (0.079)	Loss 0.934 (0.576)
Epoch: [36][120/200]	Time 0.383 (0.471)	Data 0.001 (0.078)	Loss 0.805 (0.594)
Epoch: [36][140/200]	Time 0.381 (0.472)	Data 0.001 (0.077)	Loss 0.778 (0.614)
Epoch: [36][160/200]	Time 0.383 (0.471)	Data 0.001 (0.077)	Loss 0.497 (0.625)
Epoch: [36][180/200]	Time 0.383 (0.470)	Data 0.001 (0.076)	Loss 0.680 (0.630)
Epoch: [36][200/200]	Time 0.396 (0.470)	Data 0.001 (0.075)	Loss 0.653 (0.636)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.171)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.872472763061523
==> Statistics for epoch 37: 622 clusters
Epoch: [37][20/200]	Time 1.878 (0.503)	Data 1.447 (0.110)	Loss 0.737 (0.121)
Epoch: [37][40/200]	Time 0.388 (0.482)	Data 0.001 (0.090)	Loss 0.560 (0.400)
Epoch: [37][60/200]	Time 0.378 (0.473)	Data 0.001 (0.081)	Loss 0.639 (0.496)
Epoch: [37][80/200]	Time 0.377 (0.469)	Data 0.001 (0.078)	Loss 0.622 (0.542)
Epoch: [37][100/200]	Time 0.378 (0.465)	Data 0.001 (0.075)	Loss 0.796 (0.578)
Epoch: [37][120/200]	Time 0.380 (0.463)	Data 0.001 (0.073)	Loss 0.687 (0.600)
Epoch: [37][140/200]	Time 0.379 (0.461)	Data 0.001 (0.071)	Loss 0.600 (0.614)
Epoch: [37][160/200]	Time 0.476 (0.461)	Data 0.001 (0.071)	Loss 0.789 (0.625)
Epoch: [37][180/200]	Time 0.379 (0.460)	Data 0.001 (0.070)	Loss 0.705 (0.637)
Epoch: [37][200/200]	Time 0.472 (0.459)	Data 0.001 (0.070)	Loss 0.782 (0.651)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.166)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.03859043121338
==> Statistics for epoch 38: 622 clusters
Epoch: [38][20/200]	Time 1.679 (0.493)	Data 1.259 (0.101)	Loss 0.497 (0.112)
Epoch: [38][40/200]	Time 0.379 (0.475)	Data 0.001 (0.084)	Loss 0.619 (0.394)
Epoch: [38][60/200]	Time 0.382 (0.470)	Data 0.001 (0.080)	Loss 0.565 (0.500)
Epoch: [38][80/200]	Time 0.379 (0.466)	Data 0.001 (0.075)	Loss 0.790 (0.547)
Epoch: [38][100/200]	Time 0.378 (0.464)	Data 0.001 (0.073)	Loss 1.150 (0.577)
Epoch: [38][120/200]	Time 0.378 (0.462)	Data 0.001 (0.071)	Loss 0.451 (0.614)
Epoch: [38][140/200]	Time 0.381 (0.463)	Data 0.001 (0.072)	Loss 0.760 (0.627)
Epoch: [38][160/200]	Time 0.382 (0.463)	Data 0.001 (0.071)	Loss 0.736 (0.639)
Epoch: [38][180/200]	Time 0.380 (0.462)	Data 0.001 (0.070)	Loss 0.769 (0.649)
Epoch: [38][200/200]	Time 0.386 (0.461)	Data 0.001 (0.070)	Loss 0.564 (0.659)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.167)	Data 0.000 (0.021)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.310471057891846
==> Statistics for epoch 39: 621 clusters
Epoch: [39][20/200]	Time 1.668 (0.496)	Data 1.221 (0.096)	Loss 0.860 (0.130)
Epoch: [39][40/200]	Time 0.380 (0.477)	Data 0.001 (0.084)	Loss 0.590 (0.403)
Epoch: [39][60/200]	Time 0.379 (0.468)	Data 0.001 (0.076)	Loss 0.647 (0.509)
Epoch: [39][80/200]	Time 0.382 (0.467)	Data 0.001 (0.076)	Loss 0.699 (0.557)
Epoch: [39][100/200]	Time 0.377 (0.465)	Data 0.001 (0.073)	Loss 0.639 (0.592)
Epoch: [39][120/200]	Time 0.380 (0.463)	Data 0.001 (0.072)	Loss 0.956 (0.618)
Epoch: [39][140/200]	Time 0.383 (0.462)	Data 0.001 (0.071)	Loss 0.772 (0.625)
Epoch: [39][160/200]	Time 0.386 (0.462)	Data 0.001 (0.071)	Loss 0.862 (0.630)
Epoch: [39][180/200]	Time 0.381 (0.462)	Data 0.001 (0.071)	Loss 0.684 (0.635)
Epoch: [39][200/200]	Time 0.383 (0.462)	Data 0.001 (0.071)	Loss 0.978 (0.646)
Extract Features: [50/76]	Time 0.138 (0.172)	Data 0.000 (0.024)	
Mean AP: 93.1%

 * Finished epoch  39  model mAP: 93.1%  best: 93.1% *

==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.175)	Data 0.000 (0.027)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.40622568130493
==> Statistics for epoch 40: 622 clusters
Epoch: [40][20/200]	Time 1.646 (0.492)	Data 1.211 (0.105)	Loss 0.689 (0.126)
Epoch: [40][40/200]	Time 0.385 (0.475)	Data 0.001 (0.085)	Loss 0.547 (0.395)
Epoch: [40][60/200]	Time 0.380 (0.470)	Data 0.001 (0.080)	Loss 0.646 (0.504)
Epoch: [40][80/200]	Time 0.376 (0.467)	Data 0.001 (0.077)	Loss 0.671 (0.557)
Epoch: [40][100/200]	Time 0.380 (0.465)	Data 0.001 (0.075)	Loss 0.987 (0.588)
Epoch: [40][120/200]	Time 0.465 (0.464)	Data 0.001 (0.073)	Loss 0.660 (0.602)
Epoch: [40][140/200]	Time 0.378 (0.461)	Data 0.001 (0.071)	Loss 0.723 (0.614)
Epoch: [40][160/200]	Time 0.384 (0.460)	Data 0.001 (0.070)	Loss 0.872 (0.616)
Epoch: [40][180/200]	Time 0.379 (0.460)	Data 0.001 (0.070)	Loss 0.613 (0.623)
Epoch: [40][200/200]	Time 0.380 (0.459)	Data 0.001 (0.069)	Loss 0.806 (0.637)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.167)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.153969287872314
==> Statistics for epoch 41: 619 clusters
Epoch: [41][20/200]	Time 1.748 (0.497)	Data 1.321 (0.106)	Loss 0.578 (0.116)
Epoch: [41][40/200]	Time 0.386 (0.477)	Data 0.001 (0.084)	Loss 0.635 (0.392)
Epoch: [41][60/200]	Time 0.381 (0.469)	Data 0.001 (0.076)	Loss 0.841 (0.492)
Epoch: [41][80/200]	Time 0.385 (0.465)	Data 0.001 (0.073)	Loss 0.909 (0.549)
Epoch: [41][100/200]	Time 0.381 (0.463)	Data 0.001 (0.071)	Loss 0.608 (0.579)
Epoch: [41][120/200]	Time 0.380 (0.462)	Data 0.001 (0.070)	Loss 0.542 (0.588)
Epoch: [41][140/200]	Time 0.379 (0.462)	Data 0.001 (0.069)	Loss 0.661 (0.610)
Epoch: [41][160/200]	Time 0.380 (0.462)	Data 0.001 (0.069)	Loss 0.605 (0.624)
Epoch: [41][180/200]	Time 0.382 (0.462)	Data 0.001 (0.069)	Loss 0.655 (0.631)
Epoch: [41][200/200]	Time 0.383 (0.461)	Data 0.001 (0.068)	Loss 0.808 (0.635)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.170)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.647926092147827
==> Statistics for epoch 42: 620 clusters
Epoch: [42][20/200]	Time 1.639 (0.492)	Data 1.179 (0.096)	Loss 0.751 (0.123)
Epoch: [42][40/200]	Time 0.389 (0.474)	Data 0.001 (0.081)	Loss 0.680 (0.381)
Epoch: [42][60/200]	Time 0.495 (0.469)	Data 0.001 (0.075)	Loss 0.588 (0.490)
Epoch: [42][80/200]	Time 0.377 (0.466)	Data 0.001 (0.072)	Loss 0.601 (0.558)
Epoch: [42][100/200]	Time 0.385 (0.465)	Data 0.001 (0.071)	Loss 0.711 (0.587)
Epoch: [42][120/200]	Time 0.380 (0.464)	Data 0.001 (0.070)	Loss 0.539 (0.605)
Epoch: [42][140/200]	Time 0.379 (0.465)	Data 0.001 (0.071)	Loss 0.620 (0.614)
Epoch: [42][160/200]	Time 0.401 (0.466)	Data 0.001 (0.072)	Loss 0.481 (0.626)
Epoch: [42][180/200]	Time 0.381 (0.465)	Data 0.001 (0.072)	Loss 0.674 (0.638)
Epoch: [42][200/200]	Time 0.384 (0.465)	Data 0.001 (0.071)	Loss 0.720 (0.648)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.139 (0.171)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.835617542266846
==> Statistics for epoch 43: 621 clusters
Epoch: [43][20/200]	Time 1.817 (0.498)	Data 1.277 (0.103)	Loss 0.720 (0.129)
Epoch: [43][40/200]	Time 0.392 (0.475)	Data 0.001 (0.081)	Loss 0.512 (0.389)
Epoch: [43][60/200]	Time 0.378 (0.472)	Data 0.001 (0.079)	Loss 0.527 (0.477)
Epoch: [43][80/200]	Time 0.382 (0.468)	Data 0.001 (0.075)	Loss 0.960 (0.533)
Epoch: [43][100/200]	Time 0.396 (0.466)	Data 0.001 (0.072)	Loss 0.628 (0.560)
Epoch: [43][120/200]	Time 0.485 (0.463)	Data 0.001 (0.070)	Loss 0.707 (0.579)
Epoch: [43][140/200]	Time 0.383 (0.462)	Data 0.001 (0.070)	Loss 0.591 (0.593)
Epoch: [43][160/200]	Time 0.383 (0.462)	Data 0.001 (0.069)	Loss 0.539 (0.604)
Epoch: [43][180/200]	Time 0.384 (0.461)	Data 0.001 (0.069)	Loss 0.524 (0.614)
Epoch: [43][200/200]	Time 0.380 (0.461)	Data 0.001 (0.068)	Loss 0.670 (0.629)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.169)	Data 0.000 (0.022)	
Computing jaccard distance...
Jaccard distance computing time cost: 20.221771478652954
==> Statistics for epoch 44: 620 clusters
Epoch: [44][20/200]	Time 1.647 (0.500)	Data 1.206 (0.096)	Loss 0.607 (0.135)
Epoch: [44][40/200]	Time 0.383 (0.482)	Data 0.001 (0.087)	Loss 0.393 (0.417)
Epoch: [44][60/200]	Time 0.379 (0.472)	Data 0.001 (0.077)	Loss 0.616 (0.508)
Epoch: [44][80/200]	Time 0.384 (0.467)	Data 0.001 (0.074)	Loss 0.819 (0.576)
Epoch: [44][100/200]	Time 0.387 (0.465)	Data 0.001 (0.072)	Loss 0.804 (0.602)
Epoch: [44][120/200]	Time 0.487 (0.464)	Data 0.001 (0.070)	Loss 1.090 (0.623)
Epoch: [44][140/200]	Time 0.382 (0.462)	Data 0.001 (0.069)	Loss 0.638 (0.632)
Epoch: [44][160/200]	Time 0.382 (0.462)	Data 0.001 (0.068)	Loss 0.731 (0.637)
Epoch: [44][180/200]	Time 0.495 (0.462)	Data 0.001 (0.068)	Loss 0.780 (0.645)
Epoch: [44][200/200]	Time 0.387 (0.462)	Data 0.001 (0.068)	Loss 0.767 (0.657)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.295 (0.173)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.790787935256958
==> Statistics for epoch 45: 621 clusters
Epoch: [45][20/200]	Time 1.652 (0.491)	Data 1.234 (0.097)	Loss 0.682 (0.123)
Epoch: [45][40/200]	Time 0.504 (0.479)	Data 0.001 (0.082)	Loss 0.495 (0.385)
Epoch: [45][60/200]	Time 0.381 (0.474)	Data 0.001 (0.078)	Loss 0.709 (0.493)
Epoch: [45][80/200]	Time 0.380 (0.470)	Data 0.001 (0.076)	Loss 0.954 (0.557)
Epoch: [45][100/200]	Time 0.383 (0.471)	Data 0.001 (0.076)	Loss 0.637 (0.589)
Epoch: [45][120/200]	Time 0.381 (0.469)	Data 0.001 (0.076)	Loss 0.612 (0.608)
Epoch: [45][140/200]	Time 0.379 (0.468)	Data 0.001 (0.075)	Loss 0.926 (0.630)
Epoch: [45][160/200]	Time 0.378 (0.468)	Data 0.001 (0.075)	Loss 0.683 (0.642)
Epoch: [45][180/200]	Time 0.380 (0.468)	Data 0.001 (0.075)	Loss 0.683 (0.649)
Epoch: [45][200/200]	Time 0.376 (0.468)	Data 0.001 (0.075)	Loss 0.830 (0.657)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.137 (0.174)	Data 0.000 (0.026)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.825927019119263
==> Statistics for epoch 46: 622 clusters
Epoch: [46][20/200]	Time 1.932 (0.504)	Data 1.519 (0.112)	Loss 0.827 (0.139)
Epoch: [46][40/200]	Time 0.394 (0.484)	Data 0.001 (0.092)	Loss 0.389 (0.405)
Epoch: [46][60/200]	Time 0.386 (0.477)	Data 0.001 (0.085)	Loss 0.649 (0.503)
Epoch: [46][80/200]	Time 0.380 (0.475)	Data 0.001 (0.083)	Loss 0.689 (0.555)
Epoch: [46][100/200]	Time 0.383 (0.473)	Data 0.001 (0.081)	Loss 0.731 (0.583)
Epoch: [46][120/200]	Time 0.384 (0.472)	Data 0.001 (0.080)	Loss 0.719 (0.603)
Epoch: [46][140/200]	Time 0.382 (0.470)	Data 0.001 (0.078)	Loss 0.634 (0.617)
Epoch: [46][160/200]	Time 0.387 (0.469)	Data 0.001 (0.077)	Loss 0.953 (0.629)
Epoch: [46][180/200]	Time 0.382 (0.469)	Data 0.001 (0.077)	Loss 0.807 (0.646)
Epoch: [46][200/200]	Time 0.381 (0.468)	Data 0.001 (0.077)	Loss 0.480 (0.647)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.170)	Data 0.000 (0.024)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.66496253013611
==> Statistics for epoch 47: 619 clusters
Epoch: [47][20/200]	Time 1.968 (0.507)	Data 1.554 (0.111)	Loss 0.478 (0.121)
Epoch: [47][40/200]	Time 0.514 (0.490)	Data 0.001 (0.093)	Loss 0.818 (0.417)
Epoch: [47][60/200]	Time 0.379 (0.482)	Data 0.001 (0.086)	Loss 0.457 (0.499)
Epoch: [47][80/200]	Time 0.383 (0.478)	Data 0.001 (0.083)	Loss 0.596 (0.558)
Epoch: [47][100/200]	Time 0.383 (0.474)	Data 0.002 (0.081)	Loss 0.686 (0.584)
Epoch: [47][120/200]	Time 0.383 (0.474)	Data 0.002 (0.080)	Loss 0.664 (0.600)
Epoch: [47][140/200]	Time 0.384 (0.472)	Data 0.002 (0.078)	Loss 0.641 (0.611)
Epoch: [47][160/200]	Time 0.381 (0.471)	Data 0.001 (0.077)	Loss 0.784 (0.623)
Epoch: [47][180/200]	Time 0.389 (0.471)	Data 0.001 (0.076)	Loss 0.808 (0.629)
Epoch: [47][200/200]	Time 0.382 (0.471)	Data 0.001 (0.077)	Loss 0.527 (0.637)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.172)	Data 0.000 (0.025)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.506219148635864
==> Statistics for epoch 48: 622 clusters
Epoch: [48][20/200]	Time 1.914 (0.502)	Data 1.468 (0.110)	Loss 0.528 (0.117)
Epoch: [48][40/200]	Time 0.379 (0.482)	Data 0.001 (0.094)	Loss 0.556 (0.403)
Epoch: [48][60/200]	Time 0.383 (0.478)	Data 0.001 (0.089)	Loss 0.699 (0.497)
Epoch: [48][80/200]	Time 0.380 (0.474)	Data 0.001 (0.084)	Loss 0.804 (0.549)
Epoch: [48][100/200]	Time 0.382 (0.470)	Data 0.001 (0.079)	Loss 0.683 (0.586)
Epoch: [48][120/200]	Time 0.378 (0.467)	Data 0.001 (0.077)	Loss 0.746 (0.607)
Epoch: [48][140/200]	Time 0.381 (0.466)	Data 0.001 (0.075)	Loss 0.560 (0.623)
Epoch: [48][160/200]	Time 0.379 (0.464)	Data 0.001 (0.074)	Loss 0.629 (0.632)
Epoch: [48][180/200]	Time 0.456 (0.464)	Data 0.001 (0.073)	Loss 0.622 (0.639)
Epoch: [48][200/200]	Time 0.377 (0.464)	Data 0.001 (0.073)	Loss 1.015 (0.640)
==> Create pseudo labels for unlabeled data
Extract Features: [50/51]	Time 0.138 (0.169)	Data 0.000 (0.023)	
Computing jaccard distance...
Jaccard distance computing time cost: 19.725953102111816
==> Statistics for epoch 49: 618 clusters
Epoch: [49][20/200]	Time 1.763 (0.500)	Data 1.365 (0.108)	Loss 1.201 (0.142)
Epoch: [49][40/200]	Time 0.496 (0.481)	Data 0.001 (0.087)	Loss 0.750 (0.386)
Epoch: [49][60/200]	Time 0.385 (0.472)	Data 0.001 (0.080)	Loss 0.636 (0.487)
Epoch: [49][80/200]	Time 0.379 (0.468)	Data 0.001 (0.076)	Loss 0.807 (0.526)
Epoch: [49][100/200]	Time 0.383 (0.465)	Data 0.001 (0.074)	Loss 0.433 (0.564)
Epoch: [49][120/200]	Time 0.388 (0.466)	Data 0.001 (0.075)	Loss 0.911 (0.593)
Epoch: [49][140/200]	Time 0.382 (0.464)	Data 0.001 (0.074)	Loss 0.719 (0.615)
Epoch: [49][160/200]	Time 0.380 (0.463)	Data 0.001 (0.072)	Loss 0.465 (0.620)
Epoch: [49][180/200]	Time 0.380 (0.464)	Data 0.001 (0.073)	Loss 0.709 (0.631)
Epoch: [49][200/200]	Time 0.378 (0.463)	Data 0.001 (0.072)	Loss 0.677 (0.640)
Extract Features: [50/76]	Time 0.138 (0.169)	Data 0.000 (0.022)	
Mean AP: 93.1%

 * Finished epoch  49  model mAP: 93.1%  best: 93.1%

==> Test with the best model:
=> Loaded checkpoint 'log/msmt2market/resnet101_cion/model_best.pth.tar'
Extract Features: [50/76]	Time 0.138 (0.169)	Data 0.000 (0.023)	
Mean AP: 93.1%
CMC Scores:
  top-1          96.8%
  top-5          98.8%
  top-10         99.3%
Total running time:  1:46:36.386245
