API tools faq
paste
Guest User

Untitled

a guest
Aug 20th, 2018
127
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 167.84 KB | None | 0 0
raw download clone embed print report
  1. /home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  2. return f(*args, **kwds)
  3. /home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  4. return f(*args, **kwds)
  5. /home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  6. return f(*args, **kwds)
  7. /home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  8. return f(*args, **kwds)
  9. Log file for this run: /home/ubuntu/proj/distiller-python-3.5/examples/pruning_filters_for_efficient_convnets/logs/2018.08.20-104014/2018.08.20-104014.log
  10. ==> using cifar10 dataset
  11. => creating resnet56_cifar model for CIFAR10
  12.  
  13. --------------------------------------------------------
  14. Logging to TensorBoard - remember to execute the server:
  15. > tensorboard --logdir='./logs'
  16.  
  17. => loading checkpoint logs/2018.08.20-095346/best.pth.tar
  18. Checkpoint keys:
  19. compression_sched
  20. optimizer
  21. arch
  22. state_dict
  23. best_top1
  24. epoch
  25. best top@1: 92.780
  26. Loaded compression schedule from checkpoint (epoch 167)
  27. => loaded checkpoint 'logs/2018.08.20-095346/best.pth.tar' (epoch 167)
  28. Optimizer Type: <class 'torch.optim.sgd.SGD'>
  29. Optimizer Args: {'weight_decay': 0.0001, 'momentum': 0.9, 'nesterov': False, 'lr': 0.1, 'dampening': 0}
  30. Files already downloaded and verified
  31. Files already downloaded and verified
  32. Dataset sizes:
  33. training=45000
  34. validation=5000
  35. test=10000
  36. Reading compression schedule from: resnet56_cifar_filter_rank.yaml
  37.  
  38.  
  39. Training epoch: 45000 samples (256 per mini-batch)
  40. Epoch: [168][ 50/ 176] Overall Loss 0.038152 Objective Loss 0.038152 Top1 98.757812 Top5 99.992188 LR 0.100000 Time 0.079526
  41. Epoch: [168][ 100/ 176] Overall Loss 0.115109 Objective Loss 0.115109 Top1 96.281250 Top5 99.968750 LR 0.100000 Time 0.077907
  42. Epoch: [168][ 150/ 176] Overall Loss 0.128959 Objective Loss 0.128959 Top1 95.651042 Top5 99.958333 LR 0.100000 Time 0.077437
  43.  
  44. Parameters:
  45. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  46. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  47. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  48. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.37209 | -0.00599 | 0.19940 |
  49. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11753 | -0.00660 | 0.06035 |
  50. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11910 | -0.00340 | 0.08085 |
  51. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11628 | -0.00508 | 0.07953 |
  52. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11765 | -0.00275 | 0.08621 |
  53. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11045 | -0.01236 | 0.07486 |
  54. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10654 | -0.00518 | 0.07479 |
  55. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10464 | -0.00610 | 0.06958 |
  56. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09180 | -0.00657 | 0.06176 |
  57. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10429 | -0.00683 | 0.07323 |
  58. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08431 | -0.00168 | 0.06119 |
  59. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09971 | -0.01061 | 0.07279 |
  60. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07851 | 0.00485 | 0.05603 |
  61. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09001 | -0.00868 | 0.06842 |
  62. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07469 | -0.00002 | 0.05377 |
  63. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09118 | -0.00812 | 0.06851 |
  64. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07320 | 0.00149 | 0.05518 |
  65. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09233 | -0.00916 | 0.06950 |
  66. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06607 | -0.00013 | 0.04941 |
  67. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12532 | 0.00297 | 0.09697 |
  68. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10418 | -0.00352 | 0.08068 |
  69. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.24512 | -0.00880 | 0.18041 |
  70. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08599 | -0.00361 | 0.06460 |
  71. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07525 | -0.00509 | 0.05793 |
  72. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07254 | -0.00753 | 0.05675 |
  73. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06327 | -0.00467 | 0.04914 |
  74. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06816 | -0.00788 | 0.05241 |
  75. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05744 | -0.00226 | 0.04299 |
  76. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06604 | -0.01036 | 0.05053 |
  77. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05368 | -0.00112 | 0.04064 |
  78. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06384 | -0.00563 | 0.04759 |
  79. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05186 | -0.00130 | 0.03821 |
  80. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05955 | -0.00636 | 0.04637 |
  81. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04712 | -0.00093 | 0.03531 |
  82. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05832 | -0.00614 | 0.04455 |
  83. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04683 | -0.00187 | 0.03472 |
  84. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05585 | -0.00525 | 0.04290 |
  85. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04362 | -0.00067 | 0.03264 |
  86. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08946 | -0.00452 | 0.07020 |
  87. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07883 | -0.00094 | 0.06151 |
  88. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.15204 | -0.00962 | 0.11756 |
  89. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06063 | -0.00348 | 0.04750 |
  90. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05644 | -0.00740 | 0.04406 |
  91. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05830 | -0.00252 | 0.04530 |
  92. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05192 | -0.00647 | 0.04044 |
  93. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06491 | -0.00515 | 0.05108 |
  94. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05202 | -0.00608 | 0.04046 |
  95. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06418 | -0.00253 | 0.05012 |
  96. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04405 | 0.00034 | 0.03206 |
  97. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03301 | -0.00250 | 0.02484 |
  98. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02458 | -0.00243 | 0.01756 |
  99. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03085 | -0.00143 | 0.02312 |
  100. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02228 | -0.00067 | 0.01507 |
  101. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02213 | -0.00063 | 0.01645 |
  102. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01730 | -0.00325 | 0.01238 |
  103. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03119 | -0.00095 | 0.02285 |
  104. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02215 | -0.00056 | 0.01418 |
  105. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44307 | -0.00003 | 0.32149 |
  106. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  107. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  108. Total sparsity: 0.00
  109.  
  110. --- validate (epoch=168)-----------
  111. 5000 samples (256 per mini-batch)
  112. ==> Top1: 88.000 Top5: 99.480 Loss: 0.423
  113.  
  114. ==> Best validation Top1: 88.000 Epoch: 168
  115. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  116.  
  117.  
  118. Training epoch: 45000 samples (256 per mini-batch)
  119. Epoch: [169][ 50/ 176] Overall Loss 0.100450 Objective Loss 0.100450 Top1 96.421875 Top5 99.945312 LR 0.100000 Time 0.077625
  120. Epoch: [169][ 100/ 176] Overall Loss 0.106200 Objective Loss 0.106200 Top1 96.269531 Top5 99.953125 LR 0.100000 Time 0.076929
  121. Epoch: [169][ 150/ 176] Overall Loss 0.108091 Objective Loss 0.108091 Top1 96.208333 Top5 99.955729 LR 0.100000 Time 0.076702
  122.  
  123. Parameters:
  124. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  125. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  126. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  127. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.37011 | -0.00633 | 0.19834 |
  128. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11733 | -0.00626 | 0.06036 |
  129. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11879 | -0.00464 | 0.08086 |
  130. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11592 | -0.00467 | 0.07917 |
  131. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11731 | -0.00199 | 0.08586 |
  132. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11012 | -0.01255 | 0.07459 |
  133. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10626 | -0.00477 | 0.07467 |
  134. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10453 | -0.00629 | 0.06949 |
  135. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09164 | -0.00653 | 0.06174 |
  136. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10409 | -0.00589 | 0.07340 |
  137. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08401 | -0.00229 | 0.06132 |
  138. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09966 | -0.01045 | 0.07297 |
  139. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07844 | 0.00447 | 0.05576 |
  140. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08999 | -0.00817 | 0.06837 |
  141. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07466 | 0.00074 | 0.05380 |
  142. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09152 | -0.00805 | 0.06875 |
  143. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07334 | 0.00024 | 0.05528 |
  144. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09264 | -0.00781 | 0.06974 |
  145. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06609 | -0.00107 | 0.04936 |
  146. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12495 | 0.00321 | 0.09671 |
  147. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10395 | -0.00338 | 0.08032 |
  148. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.24412 | -0.01002 | 0.17941 |
  149. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08579 | -0.00382 | 0.06439 |
  150. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07516 | -0.00495 | 0.05793 |
  151. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07255 | -0.00733 | 0.05671 |
  152. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06325 | -0.00479 | 0.04920 |
  153. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06825 | -0.00788 | 0.05248 |
  154. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05752 | -0.00195 | 0.04301 |
  155. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06603 | -0.01021 | 0.05054 |
  156. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05370 | -0.00104 | 0.04075 |
  157. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06396 | -0.00572 | 0.04770 |
  158. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05189 | -0.00103 | 0.03827 |
  159. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05958 | -0.00636 | 0.04647 |
  160. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04709 | -0.00101 | 0.03537 |
  161. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05841 | -0.00604 | 0.04462 |
  162. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04684 | -0.00164 | 0.03469 |
  163. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05597 | -0.00532 | 0.04308 |
  164. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04370 | -0.00046 | 0.03281 |
  165. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08927 | -0.00474 | 0.07007 |
  166. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07874 | -0.00097 | 0.06149 |
  167. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.15144 | -0.00907 | 0.11693 |
  168. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06068 | -0.00333 | 0.04755 |
  169. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05650 | -0.00738 | 0.04411 |
  170. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05842 | -0.00245 | 0.04544 |
  171. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05203 | -0.00633 | 0.04054 |
  172. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06498 | -0.00496 | 0.05110 |
  173. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05206 | -0.00609 | 0.04056 |
  174. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06427 | -0.00237 | 0.05018 |
  175. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04405 | 0.00045 | 0.03211 |
  176. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03311 | -0.00243 | 0.02492 |
  177. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02464 | -0.00231 | 0.01759 |
  178. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03089 | -0.00137 | 0.02316 |
  179. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02229 | -0.00067 | 0.01511 |
  180. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02216 | -0.00076 | 0.01646 |
  181. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01730 | -0.00333 | 0.01239 |
  182. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03124 | -0.00105 | 0.02291 |
  183. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02221 | -0.00056 | 0.01428 |
  184. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44385 | -0.00003 | 0.32178 |
  185. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  186. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  187. Total sparsity: 0.00
  188.  
  189. --- validate (epoch=169)-----------
  190. 5000 samples (256 per mini-batch)
  191. ==> Top1: 88.140 Top5: 99.580 Loss: 0.405
  192.  
  193. ==> Best validation Top1: 88.140 Epoch: 169
  194. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  195.  
  196.  
  197. Training epoch: 45000 samples (256 per mini-batch)
  198. Epoch: [170][ 50/ 176] Overall Loss 0.104376 Objective Loss 0.104376 Top1 96.343750 Top5 99.976562 LR 0.100000 Time 0.077781
  199. Epoch: [170][ 100/ 176] Overall Loss 0.107009 Objective Loss 0.107009 Top1 96.273438 Top5 99.980469 LR 0.100000 Time 0.077099
  200. Epoch: [170][ 150/ 176] Overall Loss 0.106209 Objective Loss 0.106209 Top1 96.291667 Top5 99.981771 LR 0.100000 Time 0.076886
  201.  
  202. Parameters:
  203. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  204. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  205. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  206. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36878 | -0.00737 | 0.19743 |
  207. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11706 | -0.00582 | 0.06033 |
  208. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11866 | -0.00459 | 0.08051 |
  209. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11571 | -0.00521 | 0.07896 |
  210. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11705 | -0.00239 | 0.08587 |
  211. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11012 | -0.01179 | 0.07462 |
  212. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10607 | -0.00547 | 0.07438 |
  213. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10445 | -0.00559 | 0.06908 |
  214. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09151 | -0.00656 | 0.06148 |
  215. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10377 | -0.00665 | 0.07355 |
  216. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08380 | -0.00220 | 0.06123 |
  217. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09956 | -0.01078 | 0.07295 |
  218. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07833 | 0.00489 | 0.05566 |
  219. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08980 | -0.00804 | 0.06809 |
  220. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07452 | -0.00002 | 0.05352 |
  221. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09183 | -0.00820 | 0.06904 |
  222. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07351 | 0.00020 | 0.05541 |
  223. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09245 | -0.00894 | 0.06968 |
  224. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06603 | -0.00109 | 0.04904 |
  225. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12462 | 0.00353 | 0.09661 |
  226. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10379 | -0.00307 | 0.08029 |
  227. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.24334 | -0.00990 | 0.17816 |
  228. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08571 | -0.00395 | 0.06443 |
  229. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07514 | -0.00464 | 0.05792 |
  230. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07266 | -0.00714 | 0.05676 |
  231. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06334 | -0.00457 | 0.04936 |
  232. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06838 | -0.00779 | 0.05274 |
  233. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05758 | -0.00198 | 0.04308 |
  234. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06609 | -0.01041 | 0.05066 |
  235. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05376 | -0.00113 | 0.04076 |
  236. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06411 | -0.00578 | 0.04780 |
  237. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05191 | -0.00092 | 0.03826 |
  238. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05971 | -0.00658 | 0.04662 |
  239. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04719 | -0.00068 | 0.03550 |
  240. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05846 | -0.00660 | 0.04464 |
  241. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04686 | -0.00174 | 0.03466 |
  242. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05607 | -0.00525 | 0.04312 |
  243. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04377 | -0.00027 | 0.03282 |
  244. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08913 | -0.00434 | 0.06994 |
  245. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07869 | -0.00095 | 0.06145 |
  246. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.15089 | -0.00898 | 0.11668 |
  247. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06076 | -0.00319 | 0.04755 |
  248. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05655 | -0.00743 | 0.04417 |
  249. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05857 | -0.00257 | 0.04557 |
  250. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05214 | -0.00637 | 0.04066 |
  251. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06506 | -0.00491 | 0.05121 |
  252. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05210 | -0.00610 | 0.04059 |
  253. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06434 | -0.00234 | 0.05023 |
  254. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04405 | 0.00026 | 0.03213 |
  255. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03322 | -0.00258 | 0.02500 |
  256. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02469 | -0.00230 | 0.01766 |
  257. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03093 | -0.00135 | 0.02319 |
  258. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02229 | -0.00057 | 0.01512 |
  259. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02219 | -0.00078 | 0.01649 |
  260. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01732 | -0.00324 | 0.01237 |
  261. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03125 | -0.00097 | 0.02294 |
  262. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02224 | -0.00057 | 0.01435 |
  263. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44304 | -0.00003 | 0.32169 |
  264. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  265. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  266. Total sparsity: 0.00
  267.  
  268. --- validate (epoch=170)-----------
  269. 5000 samples (256 per mini-batch)
  270. ==> Top1: 89.780 Top5: 99.540 Loss: 0.342
  271.  
  272. ==> Best validation Top1: 89.780 Epoch: 170
  273. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  274.  
  275.  
  276. Training epoch: 45000 samples (256 per mini-batch)
  277. Epoch: [171][ 50/ 176] Overall Loss 0.104089 Objective Loss 0.104089 Top1 96.492188 Top5 99.968750 LR 0.100000 Time 0.077790
  278. Epoch: [171][ 100/ 176] Overall Loss 0.107284 Objective Loss 0.107284 Top1 96.265625 Top5 99.972656 LR 0.100000 Time 0.077023
  279. Epoch: [171][ 150/ 176] Overall Loss 0.103784 Objective Loss 0.103784 Top1 96.351562 Top5 99.973958 LR 0.100000 Time 0.076757
  280.  
  281. Parameters:
  282. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  283. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  284. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  285. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36712 | -0.00402 | 0.19649 |
  286. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11670 | -0.00604 | 0.06012 |
  287. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11845 | -0.00324 | 0.08041 |
  288. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11551 | -0.00493 | 0.07896 |
  289. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11679 | -0.00191 | 0.08548 |
  290. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10961 | -0.01308 | 0.07423 |
  291. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10580 | -0.00550 | 0.07423 |
  292. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10423 | -0.00571 | 0.06929 |
  293. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09128 | -0.00668 | 0.06105 |
  294. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10350 | -0.00745 | 0.07357 |
  295. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08368 | -0.00176 | 0.06086 |
  296. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09925 | -0.01090 | 0.07284 |
  297. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07819 | 0.00489 | 0.05547 |
  298. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08966 | -0.00802 | 0.06801 |
  299. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07452 | 0.00024 | 0.05366 |
  300. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09196 | -0.00818 | 0.06938 |
  301. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07358 | 0.00000 | 0.05535 |
  302. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09231 | -0.00859 | 0.06979 |
  303. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06592 | -0.00050 | 0.04892 |
  304. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12438 | 0.00310 | 0.09651 |
  305. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10365 | -0.00336 | 0.08024 |
  306. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.24265 | -0.00880 | 0.17825 |
  307. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08551 | -0.00365 | 0.06434 |
  308. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07501 | -0.00506 | 0.05787 |
  309. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07265 | -0.00721 | 0.05675 |
  310. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06332 | -0.00456 | 0.04940 |
  311. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06849 | -0.00763 | 0.05283 |
  312. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05767 | -0.00224 | 0.04315 |
  313. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06609 | -0.01051 | 0.05071 |
  314. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05375 | -0.00141 | 0.04080 |
  315. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06424 | -0.00591 | 0.04793 |
  316. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05193 | -0.00098 | 0.03826 |
  317. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05979 | -0.00642 | 0.04664 |
  318. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04722 | -0.00053 | 0.03552 |
  319. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05856 | -0.00671 | 0.04476 |
  320. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04688 | -0.00182 | 0.03469 |
  321. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05615 | -0.00535 | 0.04315 |
  322. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04380 | -0.00026 | 0.03277 |
  323. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08901 | -0.00430 | 0.06993 |
  324. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07866 | -0.00115 | 0.06142 |
  325. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.15035 | -0.00923 | 0.11615 |
  326. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06081 | -0.00323 | 0.04761 |
  327. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05659 | -0.00747 | 0.04420 |
  328. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05872 | -0.00287 | 0.04573 |
  329. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05227 | -0.00626 | 0.04073 |
  330. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06507 | -0.00516 | 0.05125 |
  331. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05212 | -0.00617 | 0.04062 |
  332. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06441 | -0.00252 | 0.05034 |
  333. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04406 | 0.00029 | 0.03220 |
  334. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03330 | -0.00267 | 0.02510 |
  335. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02472 | -0.00233 | 0.01771 |
  336. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03100 | -0.00122 | 0.02323 |
  337. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02230 | -0.00065 | 0.01516 |
  338. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02224 | -0.00073 | 0.01654 |
  339. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01733 | -0.00322 | 0.01239 |
  340. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03125 | -0.00107 | 0.02297 |
  341. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02225 | -0.00061 | 0.01444 |
  342. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44352 | -0.00003 | 0.32201 |
  343. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  344. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  345. Total sparsity: 0.00
  346.  
  347. --- validate (epoch=171)-----------
  348. 5000 samples (256 per mini-batch)
  349. ==> Top1: 87.260 Top5: 99.440 Loss: 0.479
  350.  
  351. ==> Best validation Top1: 89.780 Epoch: 170
  352. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  353.  
  354.  
  355. Training epoch: 45000 samples (256 per mini-batch)
  356. Epoch: [172][ 50/ 176] Overall Loss 0.099976 Objective Loss 0.099976 Top1 96.554688 Top5 99.992188 LR 0.100000 Time 0.077654
  357. Epoch: [172][ 100/ 176] Overall Loss 0.102491 Objective Loss 0.102491 Top1 96.417969 Top5 99.988281 LR 0.100000 Time 0.076950
  358. Epoch: [172][ 150/ 176] Overall Loss 0.108099 Objective Loss 0.108099 Top1 96.187500 Top5 99.984375 LR 0.100000 Time 0.076761
  359.  
  360. Parameters:
  361. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  362. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  363. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  364. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36664 | -0.00767 | 0.19640 |
  365. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11665 | -0.00633 | 0.06015 |
  366. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11838 | -0.00297 | 0.08022 |
  367. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11550 | -0.00502 | 0.07896 |
  368. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11657 | -0.00318 | 0.08549 |
  369. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10941 | -0.01296 | 0.07401 |
  370. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10560 | -0.00550 | 0.07422 |
  371. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10414 | -0.00649 | 0.06952 |
  372. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09111 | -0.00681 | 0.06137 |
  373. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10336 | -0.00781 | 0.07344 |
  374. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08355 | -0.00156 | 0.06086 |
  375. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09916 | -0.01059 | 0.07278 |
  376. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07821 | 0.00479 | 0.05552 |
  377. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08990 | -0.00777 | 0.06854 |
  378. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07466 | -0.00034 | 0.05387 |
  379. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09203 | -0.00867 | 0.06934 |
  380. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07367 | -0.00039 | 0.05526 |
  381. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09226 | -0.00922 | 0.06984 |
  382. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06588 | -0.00076 | 0.04915 |
  383. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12436 | 0.00290 | 0.09643 |
  384. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10366 | -0.00331 | 0.08016 |
  385. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.24204 | -0.00919 | 0.17794 |
  386. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08537 | -0.00376 | 0.06425 |
  387. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07496 | -0.00517 | 0.05786 |
  388. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07267 | -0.00750 | 0.05687 |
  389. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06335 | -0.00480 | 0.04938 |
  390. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06871 | -0.00794 | 0.05308 |
  391. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05783 | -0.00244 | 0.04333 |
  392. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06626 | -0.01027 | 0.05088 |
  393. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05385 | -0.00142 | 0.04091 |
  394. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06439 | -0.00595 | 0.04813 |
  395. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05202 | -0.00124 | 0.03833 |
  396. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05993 | -0.00671 | 0.04682 |
  397. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04735 | -0.00100 | 0.03567 |
  398. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05869 | -0.00649 | 0.04484 |
  399. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04688 | -0.00204 | 0.03460 |
  400. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05639 | -0.00507 | 0.04330 |
  401. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04397 | -0.00032 | 0.03293 |
  402. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08898 | -0.00430 | 0.06991 |
  403. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07873 | -0.00097 | 0.06139 |
  404. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.15005 | -0.00873 | 0.11607 |
  405. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06095 | -0.00320 | 0.04770 |
  406. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05670 | -0.00749 | 0.04430 |
  407. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05897 | -0.00307 | 0.04603 |
  408. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05243 | -0.00634 | 0.04087 |
  409. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06520 | -0.00507 | 0.05132 |
  410. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05222 | -0.00620 | 0.04075 |
  411. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06458 | -0.00243 | 0.05051 |
  412. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04412 | 0.00026 | 0.03230 |
  413. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03346 | -0.00246 | 0.02521 |
  414. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02479 | -0.00239 | 0.01779 |
  415. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03112 | -0.00121 | 0.02330 |
  416. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02235 | -0.00048 | 0.01519 |
  417. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02231 | -0.00068 | 0.01658 |
  418. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01738 | -0.00321 | 0.01244 |
  419. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03137 | -0.00096 | 0.02305 |
  420. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02237 | -0.00065 | 0.01456 |
  421. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44094 | -0.00003 | 0.32048 |
  422. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  423. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  424. Total sparsity: 0.00
  425.  
  426. --- validate (epoch=172)-----------
  427. 5000 samples (256 per mini-batch)
  428. ==> Top1: 87.620 Top5: 99.560 Loss: 0.414
  429.  
  430. ==> Best validation Top1: 89.780 Epoch: 170
  431. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  432.  
  433.  
  434. Training epoch: 45000 samples (256 per mini-batch)
  435. Epoch: [173][ 50/ 176] Overall Loss 0.097313 Objective Loss 0.097313 Top1 96.593750 Top5 99.984375 LR 0.100000 Time 0.077791
  436. Epoch: [173][ 100/ 176] Overall Loss 0.097749 Objective Loss 0.097749 Top1 96.500000 Top5 99.984375 LR 0.100000 Time 0.077081
  437. Epoch: [173][ 150/ 176] Overall Loss 0.104300 Objective Loss 0.104300 Top1 96.260417 Top5 99.984375 LR 0.100000 Time 0.076819
  438.  
  439. Parameters:
  440. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  441. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  442. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  443. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36638 | -0.00662 | 0.19577 |
  444. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11648 | -0.00610 | 0.05998 |
  445. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11830 | -0.00401 | 0.08044 |
  446. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11524 | -0.00586 | 0.07884 |
  447. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11629 | -0.00298 | 0.08521 |
  448. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10920 | -0.01349 | 0.07427 |
  449. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10542 | -0.00550 | 0.07415 |
  450. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10397 | -0.00649 | 0.06966 |
  451. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09089 | -0.00722 | 0.06132 |
  452. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10329 | -0.00725 | 0.07337 |
  453. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08338 | -0.00208 | 0.06090 |
  454. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09913 | -0.01069 | 0.07252 |
  455. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07811 | 0.00534 | 0.05536 |
  456. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08987 | -0.00835 | 0.06843 |
  457. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07466 | -0.00017 | 0.05386 |
  458. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09202 | -0.00837 | 0.06941 |
  459. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07359 | 0.00069 | 0.05524 |
  460. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09225 | -0.00878 | 0.06955 |
  461. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06581 | -0.00063 | 0.04906 |
  462. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12415 | 0.00348 | 0.09626 |
  463. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10356 | -0.00342 | 0.08005 |
  464. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.24148 | -0.00835 | 0.17804 |
  465. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08525 | -0.00379 | 0.06418 |
  466. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07492 | -0.00483 | 0.05779 |
  467. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07263 | -0.00775 | 0.05690 |
  468. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06330 | -0.00491 | 0.04940 |
  469. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06869 | -0.00826 | 0.05317 |
  470. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05781 | -0.00236 | 0.04334 |
  471. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06626 | -0.01021 | 0.05089 |
  472. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05386 | -0.00126 | 0.04096 |
  473. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06444 | -0.00622 | 0.04830 |
  474. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05199 | -0.00129 | 0.03826 |
  475. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06003 | -0.00684 | 0.04688 |
  476. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04742 | -0.00103 | 0.03582 |
  477. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05871 | -0.00664 | 0.04488 |
  478. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04685 | -0.00178 | 0.03454 |
  479. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05629 | -0.00526 | 0.04325 |
  480. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04390 | -0.00049 | 0.03291 |
  481. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08885 | -0.00460 | 0.06983 |
  482. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07871 | -0.00098 | 0.06141 |
  483. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.14956 | -0.00926 | 0.11527 |
  484. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06099 | -0.00316 | 0.04780 |
  485. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05676 | -0.00748 | 0.04439 |
  486. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05911 | -0.00308 | 0.04615 |
  487. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05253 | -0.00622 | 0.04091 |
  488. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06525 | -0.00498 | 0.05135 |
  489. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05227 | -0.00614 | 0.04079 |
  490. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06467 | -0.00234 | 0.05056 |
  491. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04417 | 0.00029 | 0.03241 |
  492. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03356 | -0.00259 | 0.02531 |
  493. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02484 | -0.00239 | 0.01783 |
  494. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03122 | -0.00129 | 0.02340 |
  495. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02239 | -0.00055 | 0.01526 |
  496. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02239 | -0.00083 | 0.01665 |
  497. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01743 | -0.00322 | 0.01247 |
  498. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03147 | -0.00097 | 0.02311 |
  499. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02246 | -0.00065 | 0.01467 |
  500. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44085 | -0.00003 | 0.32064 |
  501. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  502. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  503. Total sparsity: 0.00
  504.  
  505. --- validate (epoch=173)-----------
  506. 5000 samples (256 per mini-batch)
  507. ==> Top1: 86.780 Top5: 99.480 Loss: 0.452
  508.  
  509. ==> Best validation Top1: 89.780 Epoch: 170
  510. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  511.  
  512.  
  513. Training epoch: 45000 samples (256 per mini-batch)
  514. Epoch: [174][ 50/ 176] Overall Loss 0.095903 Objective Loss 0.095903 Top1 96.656250 Top5 99.984375 LR 0.100000 Time 0.077624
  515. Epoch: [174][ 100/ 176] Overall Loss 0.096033 Objective Loss 0.096033 Top1 96.636719 Top5 99.984375 LR 0.100000 Time 0.076926
  516. Epoch: [174][ 150/ 176] Overall Loss 0.095752 Objective Loss 0.095752 Top1 96.627604 Top5 99.989583 LR 0.100000 Time 0.076709
  517.  
  518. Parameters:
  519. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  520. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  521. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  522. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36528 | -0.00649 | 0.19397 |
  523. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11615 | -0.00569 | 0.05968 |
  524. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11810 | -0.00415 | 0.08023 |
  525. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11489 | -0.00569 | 0.07841 |
  526. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11583 | -0.00396 | 0.08502 |
  527. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10893 | -0.01323 | 0.07417 |
  528. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10503 | -0.00587 | 0.07382 |
  529. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10380 | -0.00677 | 0.06971 |
  530. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09070 | -0.00761 | 0.06125 |
  531. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10314 | -0.00764 | 0.07342 |
  532. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08323 | -0.00177 | 0.06076 |
  533. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09891 | -0.01055 | 0.07235 |
  534. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07794 | 0.00541 | 0.05532 |
  535. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08960 | -0.00898 | 0.06856 |
  536. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07445 | 0.00039 | 0.05373 |
  537. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09198 | -0.00789 | 0.06933 |
  538. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07352 | 0.00055 | 0.05504 |
  539. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09192 | -0.00874 | 0.06936 |
  540. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06552 | -0.00043 | 0.04884 |
  541. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12383 | 0.00416 | 0.09615 |
  542. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10334 | -0.00353 | 0.07994 |
  543. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.24062 | -0.00710 | 0.17664 |
  544. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08504 | -0.00405 | 0.06409 |
  545. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07476 | -0.00521 | 0.05771 |
  546. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07257 | -0.00789 | 0.05682 |
  547. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06325 | -0.00481 | 0.04935 |
  548. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06861 | -0.00846 | 0.05326 |
  549. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05772 | -0.00243 | 0.04331 |
  550. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06621 | -0.01033 | 0.05090 |
  551. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05384 | -0.00132 | 0.04096 |
  552. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06446 | -0.00632 | 0.04839 |
  553. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05198 | -0.00129 | 0.03830 |
  554. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05997 | -0.00688 | 0.04680 |
  555. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04735 | -0.00107 | 0.03579 |
  556. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05882 | -0.00652 | 0.04494 |
  557. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04681 | -0.00160 | 0.03451 |
  558. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05619 | -0.00503 | 0.04314 |
  559. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04378 | -0.00050 | 0.03287 |
  560. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08866 | -0.00482 | 0.06978 |
  561. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07862 | -0.00100 | 0.06132 |
  562. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.14898 | -0.00868 | 0.11487 |
  563. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06095 | -0.00317 | 0.04779 |
  564. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05677 | -0.00746 | 0.04442 |
  565. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05922 | -0.00317 | 0.04621 |
  566. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05260 | -0.00625 | 0.04099 |
  567. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06529 | -0.00496 | 0.05143 |
  568. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05228 | -0.00610 | 0.04079 |
  569. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06468 | -0.00240 | 0.05058 |
  570. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04414 | 0.00015 | 0.03240 |
  571. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03361 | -0.00264 | 0.02537 |
  572. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02486 | -0.00229 | 0.01783 |
  573. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03127 | -0.00138 | 0.02342 |
  574. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02240 | -0.00058 | 0.01529 |
  575. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02241 | -0.00098 | 0.01667 |
  576. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01744 | -0.00319 | 0.01249 |
  577. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03149 | -0.00093 | 0.02314 |
  578. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02248 | -0.00055 | 0.01469 |
  579. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44186 | -0.00003 | 0.32136 |
  580. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  581. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  582. Total sparsity: 0.00
  583.  
  584. --- validate (epoch=174)-----------
  585. 5000 samples (256 per mini-batch)
  586. ==> Top1: 88.600 Top5: 99.460 Loss: 0.391
  587.  
  588. ==> Best validation Top1: 89.780 Epoch: 170
  589. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  590.  
  591.  
  592. Training epoch: 45000 samples (256 per mini-batch)
  593. Epoch: [175][ 50/ 176] Overall Loss 0.092835 Objective Loss 0.092835 Top1 96.789062 Top5 99.984375 LR 0.100000 Time 0.077622
  594. Epoch: [175][ 100/ 176] Overall Loss 0.096257 Objective Loss 0.096257 Top1 96.625000 Top5 99.968750 LR 0.100000 Time 0.076915
  595. Epoch: [175][ 150/ 176] Overall Loss 0.099672 Objective Loss 0.099672 Top1 96.476562 Top5 99.973958 LR 0.100000 Time 0.076683
  596.  
  597. Parameters:
  598. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  599. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  600. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  601. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36471 | -0.00699 | 0.19460 |
  602. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11583 | -0.00649 | 0.05960 |
  603. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11799 | -0.00347 | 0.07991 |
  604. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11485 | -0.00524 | 0.07850 |
  605. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11570 | -0.00304 | 0.08498 |
  606. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10882 | -0.01279 | 0.07390 |
  607. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10489 | -0.00602 | 0.07400 |
  608. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10386 | -0.00658 | 0.06954 |
  609. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09070 | -0.00792 | 0.06128 |
  610. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10306 | -0.00815 | 0.07365 |
  611. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08327 | -0.00243 | 0.06071 |
  612. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09888 | -0.00993 | 0.07206 |
  613. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07792 | 0.00498 | 0.05529 |
  614. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08968 | -0.00807 | 0.06853 |
  615. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07449 | 0.00029 | 0.05356 |
  616. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09195 | -0.00812 | 0.06943 |
  617. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07354 | -0.00015 | 0.05528 |
  618. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09198 | -0.00779 | 0.06938 |
  619. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06554 | -0.00066 | 0.04893 |
  620. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12373 | 0.00399 | 0.09613 |
  621. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10333 | -0.00312 | 0.07993 |
  622. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.23994 | -0.00982 | 0.17685 |
  623. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08502 | -0.00381 | 0.06419 |
  624. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07476 | -0.00519 | 0.05767 |
  625. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07270 | -0.00748 | 0.05696 |
  626. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06333 | -0.00455 | 0.04940 |
  627. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06879 | -0.00829 | 0.05332 |
  628. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05780 | -0.00200 | 0.04333 |
  629. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06618 | -0.01008 | 0.05084 |
  630. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05378 | -0.00114 | 0.04096 |
  631. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06462 | -0.00606 | 0.04843 |
  632. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05207 | -0.00109 | 0.03839 |
  633. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06008 | -0.00660 | 0.04695 |
  634. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04737 | -0.00110 | 0.03584 |
  635. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05897 | -0.00623 | 0.04500 |
  636. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04683 | -0.00153 | 0.03455 |
  637. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05624 | -0.00505 | 0.04325 |
  638. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04385 | -0.00037 | 0.03292 |
  639. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08861 | -0.00476 | 0.06973 |
  640. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07864 | -0.00118 | 0.06134 |
  641. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.14865 | -0.00825 | 0.11477 |
  642. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06102 | -0.00331 | 0.04789 |
  643. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05687 | -0.00743 | 0.04453 |
  644. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05934 | -0.00328 | 0.04635 |
  645. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05270 | -0.00627 | 0.04106 |
  646. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06541 | -0.00517 | 0.05156 |
  647. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05237 | -0.00613 | 0.04088 |
  648. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06479 | -0.00239 | 0.05064 |
  649. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04420 | 0.00015 | 0.03245 |
  650. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03370 | -0.00278 | 0.02547 |
  651. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02492 | -0.00231 | 0.01788 |
  652. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03138 | -0.00125 | 0.02353 |
  653. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02245 | -0.00060 | 0.01534 |
  654. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02249 | -0.00091 | 0.01674 |
  655. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01749 | -0.00314 | 0.01250 |
  656. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03156 | -0.00101 | 0.02322 |
  657. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02253 | -0.00057 | 0.01477 |
  658. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44113 | -0.00003 | 0.32136 |
  659. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  660. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  661. Total sparsity: 0.00
  662.  
  663. --- validate (epoch=175)-----------
  664. 5000 samples (256 per mini-batch)
  665. ==> Top1: 87.820 Top5: 99.640 Loss: 0.407
  666.  
  667. ==> Best validation Top1: 89.780 Epoch: 170
  668. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  669.  
  670.  
  671. Training epoch: 45000 samples (256 per mini-batch)
  672. Epoch: [176][ 50/ 176] Overall Loss 0.091942 Objective Loss 0.091942 Top1 97.000000 Top5 99.968750 LR 0.100000 Time 0.077814
  673. Epoch: [176][ 100/ 176] Overall Loss 0.091114 Objective Loss 0.091114 Top1 96.847656 Top5 99.980469 LR 0.100000 Time 0.077082
  674. Epoch: [176][ 150/ 176] Overall Loss 0.100282 Objective Loss 0.100282 Top1 96.526042 Top5 99.979167 LR 0.100000 Time 0.076807
  675.  
  676. Parameters:
  677. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  678. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  679. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  680. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36427 | -0.00521 | 0.19372 |
  681. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11561 | -0.00583 | 0.05963 |
  682. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11778 | -0.00387 | 0.07991 |
  683. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11487 | -0.00498 | 0.07860 |
  684. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11572 | -0.00290 | 0.08462 |
  685. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10873 | -0.01317 | 0.07399 |
  686. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10483 | -0.00503 | 0.07371 |
  687. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10387 | -0.00587 | 0.06925 |
  688. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09058 | -0.00759 | 0.06082 |
  689. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10310 | -0.00814 | 0.07378 |
  690. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08332 | -0.00203 | 0.06062 |
  691. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09901 | -0.00920 | 0.07241 |
  692. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07797 | 0.00479 | 0.05540 |
  693. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08956 | -0.00852 | 0.06849 |
  694. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07439 | 0.00098 | 0.05371 |
  695. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09197 | -0.00854 | 0.06935 |
  696. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07357 | -0.00033 | 0.05530 |
  697. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09193 | -0.00804 | 0.06934 |
  698. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06552 | -0.00035 | 0.04875 |
  699. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12357 | 0.00383 | 0.09595 |
  700. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10328 | -0.00272 | 0.07980 |
  701. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.23952 | -0.00868 | 0.17645 |
  702. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08496 | -0.00380 | 0.06415 |
  703. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07474 | -0.00523 | 0.05759 |
  704. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07266 | -0.00799 | 0.05704 |
  705. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06342 | -0.00450 | 0.04949 |
  706. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06874 | -0.00817 | 0.05325 |
  707. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05776 | -0.00198 | 0.04329 |
  708. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06615 | -0.01033 | 0.05091 |
  709. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05379 | -0.00089 | 0.04086 |
  710. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06474 | -0.00622 | 0.04848 |
  711. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05216 | -0.00100 | 0.03842 |
  712. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06022 | -0.00651 | 0.04713 |
  713. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04744 | -0.00120 | 0.03581 |
  714. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05915 | -0.00602 | 0.04515 |
  715. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04690 | -0.00147 | 0.03462 |
  716. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05627 | -0.00541 | 0.04338 |
  717. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04388 | -0.00040 | 0.03291 |
  718. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08859 | -0.00473 | 0.06971 |
  719. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07868 | -0.00117 | 0.06140 |
  720. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.14820 | -0.00865 | 0.11420 |
  721. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06111 | -0.00299 | 0.04793 |
  722. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05693 | -0.00760 | 0.04461 |
  723. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05946 | -0.00315 | 0.04645 |
  724. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05273 | -0.00642 | 0.04116 |
  725. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06548 | -0.00526 | 0.05161 |
  726. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05239 | -0.00608 | 0.04091 |
  727. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06485 | -0.00233 | 0.05070 |
  728. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04422 | 0.00011 | 0.03247 |
  729. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03378 | -0.00290 | 0.02559 |
  730. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02498 | -0.00238 | 0.01797 |
  731. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03144 | -0.00120 | 0.02358 |
  732. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02247 | -0.00057 | 0.01537 |
  733. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02255 | -0.00093 | 0.01677 |
  734. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01752 | -0.00319 | 0.01254 |
  735. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03164 | -0.00098 | 0.02329 |
  736. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02258 | -0.00056 | 0.01481 |
  737. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.43996 | -0.00003 | 0.32072 |
  738. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  739. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  740. Total sparsity: 0.00
  741.  
  742. --- validate (epoch=176)-----------
  743. 5000 samples (256 per mini-batch)
  744. ==> Top1: 87.240 Top5: 99.740 Loss: 0.427
  745.  
  746. ==> Best validation Top1: 89.780 Epoch: 170
  747. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  748.  
  749.  
  750. Training epoch: 45000 samples (256 per mini-batch)
  751. Epoch: [177][ 50/ 176] Overall Loss 0.089520 Objective Loss 0.089520 Top1 96.632812 Top5 99.992188 LR 0.100000 Time 0.077749
  752. Epoch: [177][ 100/ 176] Overall Loss 0.092172 Objective Loss 0.092172 Top1 96.687500 Top5 99.976562 LR 0.100000 Time 0.077027
  753. Epoch: [177][ 150/ 176] Overall Loss 0.094517 Objective Loss 0.094517 Top1 96.638021 Top5 99.976562 LR 0.100000 Time 0.076775
  754.  
  755. Parameters:
  756. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  757. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  758. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  759. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36340 | -0.00455 | 0.19248 |
  760. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11538 | -0.00562 | 0.05949 |
  761. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11753 | -0.00437 | 0.07978 |
  762. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11468 | -0.00540 | 0.07828 |
  763. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11544 | -0.00342 | 0.08440 |
  764. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10863 | -0.01210 | 0.07384 |
  765. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10450 | -0.00565 | 0.07370 |
  766. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10345 | -0.00593 | 0.06904 |
  767. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09023 | -0.00812 | 0.06077 |
  768. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10291 | -0.00764 | 0.07363 |
  769. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08319 | -0.00243 | 0.06066 |
  770. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09864 | -0.00934 | 0.07195 |
  771. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07781 | 0.00495 | 0.05528 |
  772. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08937 | -0.00868 | 0.06828 |
  773. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07425 | 0.00037 | 0.05356 |
  774. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09205 | -0.00872 | 0.06940 |
  775. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07348 | 0.00042 | 0.05527 |
  776. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09197 | -0.00829 | 0.06958 |
  777. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06557 | -0.00008 | 0.04863 |
  778. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12333 | 0.00416 | 0.09577 |
  779. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10312 | -0.00313 | 0.07972 |
  780. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.23866 | -0.01008 | 0.17661 |
  781. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08478 | -0.00377 | 0.06400 |
  782. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07461 | -0.00522 | 0.05747 |
  783. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07258 | -0.00784 | 0.05689 |
  784. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06340 | -0.00424 | 0.04944 |
  785. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06867 | -0.00813 | 0.05313 |
  786. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05764 | -0.00212 | 0.04316 |
  787. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06623 | -0.01040 | 0.05101 |
  788. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05387 | -0.00103 | 0.04101 |
  789. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06480 | -0.00611 | 0.04843 |
  790. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05214 | -0.00103 | 0.03841 |
  791. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06029 | -0.00681 | 0.04723 |
  792. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04748 | -0.00125 | 0.03588 |
  793. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05913 | -0.00615 | 0.04526 |
  794. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04681 | -0.00151 | 0.03455 |
  795. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05621 | -0.00546 | 0.04332 |
  796. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04383 | -0.00057 | 0.03291 |
  797. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08846 | -0.00497 | 0.06964 |
  798. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07864 | -0.00112 | 0.06140 |
  799. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.14771 | -0.00863 | 0.11404 |
  800. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06111 | -0.00307 | 0.04795 |
  801. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05695 | -0.00764 | 0.04462 |
  802. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05957 | -0.00302 | 0.04654 |
  803. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05279 | -0.00655 | 0.04123 |
  804. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06552 | -0.00522 | 0.05166 |
  805. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05242 | -0.00599 | 0.04092 |
  806. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06487 | -0.00231 | 0.05075 |
  807. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04420 | 0.00016 | 0.03248 |
  808. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03390 | -0.00279 | 0.02569 |
  809. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02502 | -0.00230 | 0.01801 |
  810. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03148 | -0.00129 | 0.02361 |
  811. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02247 | -0.00048 | 0.01536 |
  812. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02260 | -0.00102 | 0.01684 |
  813. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01757 | -0.00303 | 0.01257 |
  814. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03166 | -0.00101 | 0.02333 |
  815. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02261 | -0.00054 | 0.01485 |
  816. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.44036 | -0.00003 | 0.32085 |
  817. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  818. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  819. Total sparsity: 0.00
  820.  
  821. --- validate (epoch=177)-----------
  822. 5000 samples (256 per mini-batch)
  823. ==> Top1: 86.640 Top5: 99.580 Loss: 0.490
  824.  
  825. ==> Best validation Top1: 89.780 Epoch: 170
  826. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  827.  
  828.  
  829. Training epoch: 45000 samples (256 per mini-batch)
  830. Epoch: [178][ 50/ 176] Overall Loss 0.094499 Objective Loss 0.094499 Top1 96.632812 Top5 99.976562 LR 0.100000 Time 0.077708
  831. Epoch: [178][ 100/ 176] Overall Loss 0.096336 Objective Loss 0.096336 Top1 96.679688 Top5 99.980469 LR 0.100000 Time 0.076987
  832. Epoch: [178][ 150/ 176] Overall Loss 0.099729 Objective Loss 0.099729 Top1 96.533854 Top5 99.979167 LR 0.100000 Time 0.076752
  833.  
  834. Parameters:
  835. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  836. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  837. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  838. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36186 | -0.00486 | 0.19156 |
  839. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11514 | -0.00559 | 0.05922 |
  840. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11745 | -0.00326 | 0.07966 |
  841. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11466 | -0.00629 | 0.07817 |
  842. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11534 | -0.00266 | 0.08460 |
  843. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10840 | -0.01294 | 0.07377 |
  844. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10431 | -0.00564 | 0.07340 |
  845. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10332 | -0.00598 | 0.06915 |
  846. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09022 | -0.00724 | 0.06041 |
  847. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10283 | -0.00784 | 0.07360 |
  848. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08307 | -0.00203 | 0.06058 |
  849. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09854 | -0.00963 | 0.07189 |
  850. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07766 | 0.00548 | 0.05520 |
  851. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08917 | -0.00818 | 0.06813 |
  852. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07407 | 0.00102 | 0.05367 |
  853. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09224 | -0.00796 | 0.06956 |
  854. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07339 | 0.00017 | 0.05522 |
  855. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09226 | -0.00732 | 0.06966 |
  856. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06564 | -0.00047 | 0.04862 |
  857. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12324 | 0.00336 | 0.09567 |
  858. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10306 | -0.00367 | 0.07981 |
  859. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.23800 | -0.00983 | 0.17658 |
  860. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08475 | -0.00373 | 0.06384 |
  861. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07458 | -0.00516 | 0.05746 |
  862. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07263 | -0.00774 | 0.05707 |
  863. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06346 | -0.00439 | 0.04960 |
  864. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06860 | -0.00826 | 0.05319 |
  865. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05755 | -0.00213 | 0.04317 |
  866. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06624 | -0.01052 | 0.05099 |
  867. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05388 | -0.00065 | 0.04096 |
  868. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06481 | -0.00650 | 0.04846 |
  869. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05216 | -0.00090 | 0.03843 |
  870. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06033 | -0.00690 | 0.04722 |
  871. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04747 | -0.00111 | 0.03585 |
  872. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05935 | -0.00612 | 0.04541 |
  873. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04691 | -0.00139 | 0.03461 |
  874. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05634 | -0.00536 | 0.04352 |
  875. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04387 | -0.00053 | 0.03287 |
  876. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08846 | -0.00454 | 0.06954 |
  877. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07869 | -0.00104 | 0.06144 |
  878. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.14741 | -0.00767 | 0.11370 |
  879. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06121 | -0.00323 | 0.04804 |
  880. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05708 | -0.00758 | 0.04473 |
  881. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05973 | -0.00297 | 0.04667 |
  882. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05295 | -0.00645 | 0.04134 |
  883. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06563 | -0.00507 | 0.05174 |
  884. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05250 | -0.00594 | 0.04098 |
  885. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06497 | -0.00254 | 0.05087 |
  886. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04426 | 0.00026 | 0.03255 |
  887. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03404 | -0.00279 | 0.02580 |
  888. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02509 | -0.00232 | 0.01806 |
  889. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03156 | -0.00134 | 0.02370 |
  890. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02252 | -0.00048 | 0.01543 |
  891. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02267 | -0.00093 | 0.01689 |
  892. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01760 | -0.00302 | 0.01260 |
  893. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03175 | -0.00104 | 0.02343 |
  894. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02269 | -0.00054 | 0.01494 |
  895. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.43884 | -0.00003 | 0.32017 |
  896. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  897. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  898. Total sparsity: 0.00
  899.  
  900. --- validate (epoch=178)-----------
  901. 5000 samples (256 per mini-batch)
  902. ==> Top1: 86.680 Top5: 99.340 Loss: 0.462
  903.  
  904. ==> Best validation Top1: 89.780 Epoch: 170
  905. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  906.  
  907.  
  908. Training epoch: 45000 samples (256 per mini-batch)
  909. Epoch: [179][ 50/ 176] Overall Loss 0.102281 Objective Loss 0.102281 Top1 96.304688 Top5 99.984375 LR 0.100000 Time 0.077708
  910. Epoch: [179][ 100/ 176] Overall Loss 0.097671 Objective Loss 0.097671 Top1 96.519531 Top5 99.980469 LR 0.100000 Time 0.076979
  911. Epoch: [179][ 150/ 176] Overall Loss 0.100422 Objective Loss 0.100422 Top1 96.393229 Top5 99.973958 LR 0.100000 Time 0.076743
  912.  
  913. Parameters:
  914. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  915. | | Name | Shape | NNZ (dense) | NNZ (sparse) | Cols (%) | Rows (%) | Ch (%) | 2D (%) | 3D (%) | Fine (%) | Std | Mean | Abs-Mean |
  916. |----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------|
  917. | 0 | module.conv1.weight | (16, 3, 3, 3) | 432 | 432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.36177 | -0.00649 | 0.19179 |
  918. | 1 | module.layer1.0.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11498 | -0.00503 | 0.05924 |
  919. | 2 | module.layer1.0.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11726 | -0.00376 | 0.07974 |
  920. | 3 | module.layer1.1.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11471 | -0.00600 | 0.07837 |
  921. | 4 | module.layer1.1.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.11535 | -0.00383 | 0.08472 |
  922. | 5 | module.layer1.2.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10847 | -0.01247 | 0.07380 |
  923. | 6 | module.layer1.2.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10426 | -0.00616 | 0.07334 |
  924. | 7 | module.layer1.3.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10326 | -0.00562 | 0.06918 |
  925. | 8 | module.layer1.3.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09013 | -0.00750 | 0.06037 |
  926. | 9 | module.layer1.4.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10284 | -0.00813 | 0.07370 |
  927. | 10 | module.layer1.4.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08307 | -0.00230 | 0.06060 |
  928. | 11 | module.layer1.5.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09847 | -0.01050 | 0.07200 |
  929. | 12 | module.layer1.5.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07775 | 0.00528 | 0.05502 |
  930. | 13 | module.layer1.6.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08891 | -0.00828 | 0.06789 |
  931. | 14 | module.layer1.6.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07393 | 0.00030 | 0.05350 |
  932. | 15 | module.layer1.7.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09259 | -0.00762 | 0.06968 |
  933. | 16 | module.layer1.7.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07352 | -0.00034 | 0.05535 |
  934. | 17 | module.layer1.8.conv1.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.09228 | -0.00801 | 0.06977 |
  935. | 18 | module.layer1.8.conv2.weight | (16, 16, 3, 3) | 2304 | 2304 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06560 | -0.00029 | 0.04837 |
  936. | 19 | module.layer2.0.conv1.weight | (32, 16, 3, 3) | 4608 | 4608 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.12317 | 0.00391 | 0.09558 |
  937. | 20 | module.layer2.0.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.10308 | -0.00326 | 0.07979 |
  938. | 21 | module.layer2.0.downsample.0.weight | (32, 16, 1, 1) | 512 | 512 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.23748 | -0.00816 | 0.17551 |
  939. | 22 | module.layer2.1.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08473 | -0.00331 | 0.06386 |
  940. | 23 | module.layer2.1.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07461 | -0.00513 | 0.05755 |
  941. | 24 | module.layer2.2.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07256 | -0.00833 | 0.05717 |
  942. | 25 | module.layer2.2.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06346 | -0.00437 | 0.04959 |
  943. | 26 | module.layer2.3.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06867 | -0.00831 | 0.05317 |
  944. | 27 | module.layer2.3.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05754 | -0.00205 | 0.04314 |
  945. | 28 | module.layer2.4.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06630 | -0.00995 | 0.05095 |
  946. | 29 | module.layer2.4.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05386 | -0.00104 | 0.04097 |
  947. | 30 | module.layer2.5.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06494 | -0.00622 | 0.04856 |
  948. | 31 | module.layer2.5.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05220 | -0.00088 | 0.03849 |
  949. | 32 | module.layer2.6.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06047 | -0.00667 | 0.04733 |
  950. | 33 | module.layer2.6.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04749 | -0.00095 | 0.03584 |
  951. | 34 | module.layer2.7.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05955 | -0.00646 | 0.04571 |
  952. | 35 | module.layer2.7.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04700 | -0.00156 | 0.03474 |
  953. | 36 | module.layer2.8.conv1.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05645 | -0.00519 | 0.04357 |
  954. | 37 | module.layer2.8.conv2.weight | (32, 32, 3, 3) | 9216 | 9216 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04392 | -0.00051 | 0.03294 |
  955. | 38 | module.layer3.0.conv1.weight | (64, 32, 3, 3) | 18432 | 18432 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.08843 | -0.00465 | 0.06951 |
  956. | 39 | module.layer3.0.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.07872 | -0.00113 | 0.06148 |
  957. | 40 | module.layer3.0.downsample.0.weight | (64, 32, 1, 1) | 2048 | 2048 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.14702 | -0.00783 | 0.11348 |
  958. | 41 | module.layer3.1.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06124 | -0.00356 | 0.04816 |
  959. | 42 | module.layer3.1.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05713 | -0.00772 | 0.04480 |
  960. | 43 | module.layer3.2.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05983 | -0.00308 | 0.04676 |
  961. | 44 | module.layer3.2.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05304 | -0.00644 | 0.04144 |
  962. | 45 | module.layer3.3.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06570 | -0.00521 | 0.05182 |
  963. | 46 | module.layer3.3.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.05255 | -0.00598 | 0.04105 |
  964. | 47 | module.layer3.4.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.06502 | -0.00241 | 0.05093 |
  965. | 48 | module.layer3.4.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.04429 | 0.00034 | 0.03261 |
  966. | 49 | module.layer3.5.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03411 | -0.00294 | 0.02593 |
  967. | 50 | module.layer3.5.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02515 | -0.00225 | 0.01808 |
  968. | 51 | module.layer3.6.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03166 | -0.00114 | 0.02378 |
  969. | 52 | module.layer3.6.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02256 | -0.00040 | 0.01548 |
  970. | 53 | module.layer3.7.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02273 | -0.00101 | 0.01694 |
  971. | 54 | module.layer3.7.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.01762 | -0.00293 | 0.01260 |
  972. | 55 | module.layer3.8.conv1.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.03182 | -0.00079 | 0.02345 |
  973. | 56 | module.layer3.8.conv2.weight | (64, 64, 3, 3) | 36864 | 36864 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.02275 | -0.00041 | 0.01502 |
  974. | 57 | module.fc.weight | (10, 64) | 640 | 640 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.43776 | -0.00003 | 0.31952 |
  975. | 58 | Total sparsity: | - | 851504 | 851504 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 | 0.00000 |
  976. +----+-------------------------------------+----------------+---------------+----------------+------------+------------+----------+----------+----------+------------+---------+----------+------------+
  977. Total sparsity: 0.00
  978.  
  979. --- validate (epoch=179)-----------
  980. 5000 samples (256 per mini-batch)
  981. ==> Top1: 87.920 Top5: 99.380 Loss: 0.456
  982.  
  983. ==> Best validation Top1: 89.780 Epoch: 170
  984. Saving checkpoint to: logs/2018.08.20-104014/checkpoint.pth.tar
  985.  
  986.  
  987. L1RankedStructureParameterPruner - param: module.layer1.0.conv1.weight pruned=0.562 goal=0.600 (9/16)
  988. L1RankedStructureParameterPruner - param: module.layer1.1.conv1.weight pruned=0.562 goal=0.600 (9/16)
  989. L1RankedStructureParameterPruner - param: module.layer1.2.conv1.weight pruned=0.562 goal=0.600 (9/16)
  990. L1RankedStructureParameterPruner - param: module.layer1.3.conv1.weight pruned=0.562 goal=0.600 (9/16)
  991. L1RankedStructureParameterPruner - param: module.layer1.4.conv1.weight pruned=0.562 goal=0.600 (9/16)
  992. L1RankedStructureParameterPruner - param: module.layer1.5.conv1.weight pruned=0.562 goal=0.600 (9/16)
  993. L1RankedStructureParameterPruner - param: module.layer1.6.conv1.weight pruned=0.562 goal=0.600 (9/16)
  994. L1RankedStructureParameterPruner - param: module.layer1.7.conv1.weight pruned=0.562 goal=0.600 (9/16)
  995. L1RankedStructureParameterPruner - param: module.layer1.8.conv1.weight pruned=0.562 goal=0.600 (9/16)
  996. L1RankedStructureParameterPruner - param: module.layer2.1.conv1.weight pruned=0.500 goal=0.500 (16/32)
  997. L1RankedStructureParameterPruner - param: module.layer2.2.conv1.weight pruned=0.500 goal=0.500 (16/32)
  998. L1RankedStructureParameterPruner - param: module.layer2.3.conv1.weight pruned=0.500 goal=0.500 (16/32)
  999. L1RankedStructureParameterPruner - param: module.layer2.4.conv1.weight pruned=0.500 goal=0.500 (16/32)
  1000. L1RankedStructureParameterPruner - param: module.layer2.6.conv1.weight pruned=0.500 goal=0.500 (16/32)
  1001. L1RankedStructureParameterPruner - param: module.layer2.7.conv1.weight pruned=0.500 goal=0.500 (16/32)
  1002. L1RankedStructureParameterPruner - param: module.layer3.1.conv1.weight pruned=0.094 goal=0.100 (6/64)
  1003. L1RankedStructureParameterPruner - param: module.layer3.2.conv1.weight pruned=0.297 goal=0.300 (19/64)
  1004. L1RankedStructureParameterPruner - param: module.layer3.3.conv1.weight pruned=0.297 goal=0.300 (19/64)
  1005. L1RankedStructureParameterPruner - param: module.layer3.5.conv1.weight pruned=0.297 goal=0.300 (19/64)
  1006. L1RankedStructureParameterPruner - param: module.layer3.6.conv1.weight pruned=0.297 goal=0.300 (19/64)
  1007. L1RankedStructureParameterPruner - param: module.layer3.7.conv1.weight pruned=0.297 goal=0.300 (19/64)
  1008. L1RankedStructureParameterPruner - param: module.layer3.8.conv1.weight pruned=0.297 goal=0.300 (19/64)
  1009. Training epoch: 45000 samples (256 per mini-batch)
  1010. ==> using cifar10 dataset
  1011. => creating resnet56_cifar model for CIFAR10
  1012. Invoking create_thinning_recipe_filters
  1013. In tensor module.layer1.0.conv1.weight found 9/16 zero filters
  1014. In tensor module.layer1.1.conv1.weight found 9/16 zero filters
  1015. In tensor module.layer1.2.conv1.weight found 9/16 zero filters
  1016. In tensor module.layer1.3.conv1.weight found 9/16 zero filters
  1017. In tensor module.layer1.4.conv1.weight found 9/16 zero filters
  1018. In tensor module.layer1.5.conv1.weight found 9/16 zero filters
  1019. In tensor module.layer1.6.conv1.weight found 9/16 zero filters
  1020. In tensor module.layer1.7.conv1.weight found 9/16 zero filters
  1021. In tensor module.layer1.8.conv1.weight found 9/16 zero filters
  1022. In tensor module.layer2.1.conv1.weight found 16/32 zero filters
  1023. In tensor module.layer2.2.conv1.weight found 16/32 zero filters
  1024. In tensor module.layer2.3.conv1.weight found 16/32 zero filters
  1025. In tensor module.layer2.4.conv1.weight found 16/32 zero filters
  1026. In tensor module.layer2.6.conv1.weight found 16/32 zero filters
  1027. In tensor module.layer2.7.conv1.weight found 16/32 zero filters
  1028. In tensor module.layer3.1.conv1.weight found 6/64 zero filters
  1029. In tensor module.layer3.2.conv1.weight found 19/64 zero filters
  1030. In tensor module.layer3.3.conv1.weight found 19/64 zero filters
  1031. In tensor module.layer3.5.conv1.weight found 19/64 zero filters
  1032. In tensor module.layer3.6.conv1.weight found 19/64 zero filters
  1033. In tensor module.layer3.7.conv1.weight found 19/64 zero filters
  1034. In tensor module.layer3.8.conv1.weight found 19/64 zero filters
  1035. Created, applied and saved a thinning recipe
  1036. Traceback (most recent call last):
  1037. File "../classifier_compression/compress_classifier.py", line 688, in <module>
  1038. main()
  1039. File "../classifier_compression/compress_classifier.py", line 296, in main
  1040. loggers=[tflogger, pylogger], args=args)
  1041. File "../classifier_compression/compress_classifier.py", line 369, in train
  1042. output = model(input_var)
  1043. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1044. result = self.forward(*input, **kwargs)
  1045. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/parallel/data_parallel.py", line 112, in forward
  1046. return self.module(*inputs[0], **kwargs[0])
  1047. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1048. result = self.forward(*input, **kwargs)
  1049. File "/home/ubuntu/proj/distiller-python-3.5/models/cifar10/resnet_cifar.py", line 140, in forward
  1050. x = self.layer1(x)
  1051. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1052. result = self.forward(*input, **kwargs)
  1053. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/container.py", line 91, in forward
  1054. input = module(input)
  1055. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1056. result = self.forward(*input, **kwargs)
  1057. File "/home/ubuntu/proj/distiller-python-3.5/models/cifar10/resnet_cifar.py", line 71, in forward
  1058. out = self.bn1(out)
  1059. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1060. result = self.forward(*input, **kwargs)
  1061. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/batchnorm.py", line 49, in forward
  1062. self.training or not self.track_running_stats, self.momentum, self.eps)
  1063. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/functional.py", line 1194, in batch_norm
  1064. training, momentum, eps, torch.backends.cudnn.enabled
  1065. RuntimeError: running_mean should contain 7 elements not 16
  1066.  
  1067.  
  1068. Log file for this run: /home/ubuntu/proj/distiller-python-3.5/examples/pruning_filters_for_efficient_convnets/logs/2018.08.20-104014/2018.08.20-104014.log
  1069. Traceback (most recent call last):
  1070. File "../classifier_compression/compress_classifier.py", line 688, in <module>
  1071. main()
  1072. File "../classifier_compression/compress_classifier.py", line 296, in main
  1073. loggers=[tflogger, pylogger], args=args)
  1074. File "../classifier_compression/compress_classifier.py", line 369, in train
  1075. output = model(input_var)
  1076. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1077. result = self.forward(*input, **kwargs)
  1078. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/parallel/data_parallel.py", line 112, in forward
  1079. return self.module(*inputs[0], **kwargs[0])
  1080. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1081. result = self.forward(*input, **kwargs)
  1082. File "/home/ubuntu/proj/distiller-python-3.5/models/cifar10/resnet_cifar.py", line 140, in forward
  1083. x = self.layer1(x)
  1084. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1085. result = self.forward(*input, **kwargs)
  1086. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/container.py", line 91, in forward
  1087. input = module(input)
  1088. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1089. result = self.forward(*input, **kwargs)
  1090. File "/home/ubuntu/proj/distiller-python-3.5/models/cifar10/resnet_cifar.py", line 71, in forward
  1091. out = self.bn1(out)
  1092. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/module.py", line 491, in __call__
  1093. result = self.forward(*input, **kwargs)
  1094. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/modules/batchnorm.py", line 49, in forward
  1095. self.training or not self.track_running_stats, self.momentum, self.eps)
  1096. File "/home/ubuntu/conda/miniconda3/envs/distiller-python-3.5/lib/python3.5/site-packages/torch/nn/functional.py", line 1194, in batch_norm
  1097. training, momentum, eps, torch.backends.cudnn.enabled
  1098. RuntimeError: running_mean should contain 7 elements not 16
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!