text
stringlengths 5
1.13k
|
|---|
2020-02-02 11:35:39 Iteration 150 Training Loss: 2.231e-01 Loss in Target Net: 2.240e-02 |
2020-02-02 11:35:58 Iteration 200 Training Loss: 2.173e-01 Loss in Target Net: 2.326e-02 |
2020-02-02 11:36:16 Iteration 250 Training Loss: 2.104e-01 Loss in Target Net: 1.900e-02 |
2020-02-02 11:36:33 Iteration 300 Training Loss: 2.018e-01 Loss in Target Net: 2.287e-02 |
2020-02-02 11:36:52 Iteration 350 Training Loss: 2.029e-01 Loss in Target Net: 2.163e-02 |
2020-02-02 11:37:11 Iteration 400 Training Loss: 1.995e-01 Loss in Target Net: 2.197e-02 |
2020-02-02 11:37:29 Iteration 450 Training Loss: 2.034e-01 Loss in Target Net: 2.170e-02 |
2020-02-02 11:37:47 Iteration 500 Training Loss: 1.999e-01 Loss in Target Net: 2.057e-02 |
2020-02-02 11:38:05 Iteration 550 Training Loss: 1.973e-01 Loss in Target Net: 2.349e-02 |
2020-02-02 11:38:23 Iteration 600 Training Loss: 1.957e-01 Loss in Target Net: 2.181e-02 |
2020-02-02 11:38:41 Iteration 650 Training Loss: 2.002e-01 Loss in Target Net: 2.035e-02 |
2020-02-02 11:38:59 Iteration 700 Training Loss: 1.922e-01 Loss in Target Net: 2.567e-02 |
2020-02-02 11:39:17 Iteration 750 Training Loss: 1.946e-01 Loss in Target Net: 2.036e-02 |
2020-02-02 11:39:34 Iteration 800 Training Loss: 1.955e-01 Loss in Target Net: 2.302e-02 |
2020-02-02 11:39:53 Iteration 850 Training Loss: 1.918e-01 Loss in Target Net: 2.370e-02 |
2020-02-02 11:40:11 Iteration 900 Training Loss: 1.946e-01 Loss in Target Net: 2.282e-02 |
2020-02-02 11:40:28 Iteration 950 Training Loss: 1.954e-01 Loss in Target Net: 2.010e-02 |
2020-02-02 11:40:48 Iteration 1000 Training Loss: 1.905e-01 Loss in Target Net: 2.001e-02 |
2020-02-02 11:41:07 Iteration 1050 Training Loss: 1.929e-01 Loss in Target Net: 1.861e-02 |
2020-02-02 11:41:25 Iteration 1100 Training Loss: 1.922e-01 Loss in Target Net: 2.059e-02 |
2020-02-02 11:41:44 Iteration 1150 Training Loss: 1.919e-01 Loss in Target Net: 2.001e-02 |
2020-02-02 11:42:03 Iteration 1200 Training Loss: 1.917e-01 Loss in Target Net: 1.969e-02 |
2020-02-02 11:42:21 Iteration 1250 Training Loss: 1.890e-01 Loss in Target Net: 1.987e-02 |
2020-02-02 11:42:41 Iteration 1300 Training Loss: 1.941e-01 Loss in Target Net: 2.085e-02 |
2020-02-02 11:43:00 Iteration 1350 Training Loss: 1.971e-01 Loss in Target Net: 2.031e-02 |
2020-02-02 11:43:19 Iteration 1400 Training Loss: 1.928e-01 Loss in Target Net: 2.364e-02 |
2020-02-02 11:43:38 Iteration 1450 Training Loss: 1.916e-01 Loss in Target Net: 2.090e-02 |
2020-02-02 11:43:56 Iteration 1499 Training Loss: 1.901e-01 Loss in Target Net: 2.036e-02 |
Evaluating against victims networks |
DPN92 |
Using Adam for retraining |
Files already downloaded and verified |
2020-02-02 11:44:06, Epoch 0, Iteration 7, loss 0.255 (0.475), acc 88.462 (89.000) |
2020-02-02 11:45:04, Epoch 30, Iteration 7, loss 0.000 (0.000), acc 100.000 (100.000) |
Target Label: 6, Poison label: 8, Prediction:6, Target's Score:[-3.1243932, -0.763719, -2.2759798, 0.079497464, -0.36522615, -2.813989, 8.836633, -2.817194, 5.293566, -1.5606623], Poisons' Predictions:[8, 8, 8, 8, 8] |
2020-02-02 11:46:03 Epoch 59, Val iteration 0, acc 91.800 (91.800) |
2020-02-02 11:46:11 Epoch 59, Val iteration 19, acc 92.600 (92.950) |
* Prec: 92.95000076293945 |
-------- |
------SUMMARY------ |
TIME ELAPSED (mins): 9 |
TARGET INDEX: 16 |
DPN92 0 |
Namespace(chk_path='chk-black-end2end', chk_subdir='poisons', device='cuda', dset_path='datasets', end2end=True, eval_poison_path='', gpu='1', lr_decay_epoch=[30, 45], mode='mean', model_resume_path='model-chks', nearest=False, net_repeat=1, num_per_class=50, original_grad=True, poison_decay_ites=[], poison_decay_ratio=0.1, poison_epsilon=0.1, poison_ites=1500, poison_label=8, poison_lr=0.04, poison_momentum=0.9, poison_num=5, poison_opt='adam', resume_poison_ite=0, retrain_bsize=64, retrain_epochs=60, retrain_lr=0.0001, retrain_momentum=0.9, retrain_opt='adam', retrain_wd=0.0005, subs_chk_name=['ckpt-%s-4800-dp0.200-droplayer0.000-seed1226.t7', 'ckpt-%s-4800-dp0.250-droplayer0.000-seed1226.t7', 'ckpt-%s-4800-dp0.300-droplayer0.000.t7'], subs_dp=[0.2, 0.25, 0.3], subset_group=0, substitute_nets=['DPN92', 'SENet18', 'ResNet50', 'ResNeXt29_2x64d'], target_index=17, target_label=6, target_net=['DPN92'], test_chk_name='ckpt-%s-4800.t7', tol=1e-06, train_data_path='datasets/CIFAR10_TRAIN_Split.pth') |
Path: chk-black-end2end/mean/1500/17 |
Selected base image indices: [213, 225, 227, 247, 249] |
2020-02-02 11:36:55 Iteration 0 Training Loss: 1.012e+00 Loss in Target Net: 1.376e+00 |
2020-02-02 11:37:13 Iteration 50 Training Loss: 2.217e-01 Loss in Target Net: 5.175e-02 |
2020-02-02 11:37:30 Iteration 100 Training Loss: 2.089e-01 Loss in Target Net: 4.775e-02 |
2020-02-02 11:37:48 Iteration 150 Training Loss: 1.932e-01 Loss in Target Net: 4.708e-02 |
2020-02-02 11:38:05 Iteration 200 Training Loss: 1.914e-01 Loss in Target Net: 4.692e-02 |
2020-02-02 11:38:23 Iteration 250 Training Loss: 1.859e-01 Loss in Target Net: 4.544e-02 |
2020-02-02 11:38:43 Iteration 300 Training Loss: 1.800e-01 Loss in Target Net: 4.413e-02 |
2020-02-02 11:39:01 Iteration 350 Training Loss: 1.768e-01 Loss in Target Net: 5.271e-02 |
2020-02-02 11:39:19 Iteration 400 Training Loss: 1.802e-01 Loss in Target Net: 4.826e-02 |
2020-02-02 11:39:37 Iteration 450 Training Loss: 1.782e-01 Loss in Target Net: 4.869e-02 |
2020-02-02 11:39:57 Iteration 500 Training Loss: 1.790e-01 Loss in Target Net: 5.045e-02 |
2020-02-02 11:40:15 Iteration 550 Training Loss: 1.767e-01 Loss in Target Net: 4.563e-02 |
2020-02-02 11:40:34 Iteration 600 Training Loss: 1.808e-01 Loss in Target Net: 4.156e-02 |
2020-02-02 11:40:53 Iteration 650 Training Loss: 1.771e-01 Loss in Target Net: 3.740e-02 |
2020-02-02 11:41:11 Iteration 700 Training Loss: 1.770e-01 Loss in Target Net: 3.552e-02 |
2020-02-02 11:41:30 Iteration 750 Training Loss: 1.805e-01 Loss in Target Net: 4.468e-02 |
2020-02-02 11:41:49 Iteration 800 Training Loss: 1.759e-01 Loss in Target Net: 4.270e-02 |
2020-02-02 11:42:07 Iteration 850 Training Loss: 1.725e-01 Loss in Target Net: 4.101e-02 |
2020-02-02 11:42:25 Iteration 900 Training Loss: 1.765e-01 Loss in Target Net: 4.316e-02 |
2020-02-02 11:42:44 Iteration 950 Training Loss: 1.745e-01 Loss in Target Net: 3.277e-02 |
2020-02-02 11:43:02 Iteration 1000 Training Loss: 1.714e-01 Loss in Target Net: 3.918e-02 |
2020-02-02 11:43:20 Iteration 1050 Training Loss: 1.723e-01 Loss in Target Net: 4.340e-02 |
2020-02-02 11:43:38 Iteration 1100 Training Loss: 1.718e-01 Loss in Target Net: 3.716e-02 |
2020-02-02 11:43:56 Iteration 1150 Training Loss: 1.703e-01 Loss in Target Net: 3.592e-02 |
2020-02-02 11:44:16 Iteration 1200 Training Loss: 1.713e-01 Loss in Target Net: 3.478e-02 |
2020-02-02 11:44:36 Iteration 1250 Training Loss: 1.776e-01 Loss in Target Net: 3.930e-02 |
2020-02-02 11:44:56 Iteration 1300 Training Loss: 1.722e-01 Loss in Target Net: 4.184e-02 |
2020-02-02 11:45:16 Iteration 1350 Training Loss: 1.675e-01 Loss in Target Net: 3.327e-02 |
2020-02-02 11:45:35 Iteration 1400 Training Loss: 1.743e-01 Loss in Target Net: 3.507e-02 |
2020-02-02 11:45:55 Iteration 1450 Training Loss: 1.738e-01 Loss in Target Net: 3.668e-02 |
2020-02-02 11:46:13 Iteration 1499 Training Loss: 1.683e-01 Loss in Target Net: 3.233e-02 |
Evaluating against victims networks |
DPN92 |
Using Adam for retraining |
Files already downloaded and verified |
2020-02-02 11:46:23, Epoch 0, Iteration 7, loss 0.530 (0.413), acc 80.769 (90.600) |
2020-02-02 11:47:21, Epoch 30, Iteration 7, loss 0.000 (0.000), acc 100.000 (100.000) |
Target Label: 6, Poison label: 8, Prediction:6, Target's Score:[-2.4170392, -1.3972242, -2.4566488, -1.4624834, -0.40297973, -1.3363873, 9.2487335, -2.8370438, 4.9661694, -1.4730896], Poisons' Predictions:[8, 8, 8, 8, 8] |
2020-02-02 11:48:20 Epoch 59, Val iteration 0, acc 93.800 (93.800) |
2020-02-02 11:48:28 Epoch 59, Val iteration 19, acc 93.000 (93.320) |
* Prec: 93.32000198364258 |
-------- |
------SUMMARY------ |
TIME ELAPSED (mins): 9 |
TARGET INDEX: 17 |
DPN92 0 |
Namespace(chk_path='chk-black-end2end', chk_subdir='poisons', device='cuda', dset_path='datasets', end2end=True, eval_poison_path='', gpu='1', lr_decay_epoch=[30, 45], mode='mean', model_resume_path='model-chks', nearest=False, net_repeat=1, num_per_class=50, original_grad=True, poison_decay_ites=[], poison_decay_ratio=0.1, poison_epsilon=0.1, poison_ites=1500, poison_label=8, poison_lr=0.04, poison_momentum=0.9, poison_num=5, poison_opt='adam', resume_poison_ite=0, retrain_bsize=64, retrain_epochs=60, retrain_lr=0.0001, retrain_momentum=0.9, retrain_opt='adam', retrain_wd=0.0005, subs_chk_name=['ckpt-%s-4800-dp0.200-droplayer0.000-seed1226.t7', 'ckpt-%s-4800-dp0.250-droplayer0.000-seed1226.t7', 'ckpt-%s-4800-dp0.300-droplayer0.000.t7'], subs_dp=[0.2, 0.25, 0.3], subset_group=0, substitute_nets=['DPN92', 'SENet18', 'ResNet50', 'ResNeXt29_2x64d'], target_index=17, target_label=6, target_net=['DPN92'], test_chk_name='ckpt-%s-4800.t7', tol=1e-06, train_data_path='datasets/CIFAR10_TRAIN_Split.pth') |
Path: chk-black-end2end/mean/1500/17 |
Selected base image indices: [213, 225, 227, 247, 249] |
2020-02-03 04:25:10 Iteration 0 Training Loss: 1.015e+00 Loss in Target Net: 1.410e+00 |
2020-02-03 04:25:31 Iteration 50 Training Loss: 2.294e-01 Loss in Target Net: 5.890e-02 |
2020-02-03 04:25:51 Iteration 100 Training Loss: 2.040e-01 Loss in Target Net: 4.952e-02 |
2020-02-03 04:26:31 Iteration 150 Training Loss: 1.925e-01 Loss in Target Net: 5.164e-02 |
2020-02-03 04:27:42 Iteration 200 Training Loss: 1.857e-01 Loss in Target Net: 3.915e-02 |
Subsets and Splits