ziangyan / pda.pytorch Goto Github PK

I0325 05:30:26.979458 11136 policy_attack.py:1433] victim_batch_size: 50
I0325 05:30:32.276054 11136 policy_attack.py:381] Found policy keys: OrderedDict([('policy_weight_fname', 'output/policy-smtrain-inv-mnist-carlinet-untargeted-l2/2020-09-28_13-41-55-DgrgzXkj/results/model_best.pth'), ('policy_bilinear', True), ('policy_normalization_type', 'none'), ('policy_use_tanh', False), ('policy_base_width', 32), ('policy_calibrate', True), ('policy_init_std', 0.003)]), only use OrderedDict([('init_std', 0.003), ('input_size', 0), ('weight_fname', 'output/policy-smtrain-inv-mnist-carlinet-untargeted-l2/2020-09-28_13-41-55-DgrgzXkj/results/model_best.pth'), ('calibrate', True), ('use_tanh', False), ('normalization_type', 'none')]) for dataset mnist and arch vgg13_inv
Traceback (most recent call last):
File "policy_attack.py", line 1503, in
main()
File "policy_attack.py", line 382, in main
policy = make_policy_model(args.dataset, args.policy_arch, **kwargs).train().to(device)
File "/home/pda.pytorch-master/models/init.py", line 202, in make_policy_model
assert osp.exists(weight_fname)
AssertionError
root@149587cc4c1c:/home/pda.pytorch-master#