roysubhankar / dwt-domain-adaptation Goto Github PK

In resnet50_dwt_mec_officehome.py line 524, cuda device fixed as 0.
In whitening.py line 23,24,48,51, cuda device fixed as 0.
This prohibits this code from running on non-"0" gpus.
Would be nice to make device id an argument.

hi, thanks for your contribution
I have some problem here
Traceback (most recent call last):
File "resnet50_dwt_mec_officehome.py", line 603, in
main()
File "resnet50_dwt_mec_officehome.py", line 600, in main
target_test_loader=target_testloader)
File "resnet50_dwt_mec_officehome.py", line 420, in train_infinite_collect_stats
output = model(data)
File "/home/xing/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "resnet50_dwt_mec_officehome.py", line 336, in forward
x = torch.cat((self.bns1(x_s), torch.cat((self.bnt1(x_t), self.bnt1_aug(x_t_dup)), dim=0) ), dim=0) * self.gamma1 + self.beta1
File "/home/xing/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "resnet50_dwt_mec_officehome.py", line 60, in forward
out = self.wh(x)
File "/home/xing/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "/home/xing/mengzhu/paper_code/dwt-domain-adaptation-master/utils/whitening.py", line 50, in forward
f_cov_shrinked = (1-self.eps) * f_cov + self.eps * torch.eye(self.group_size, out=torch.cuda(2).FloatTensor() if torch.cuda.is_available() else torch.FloatTensor()).repeat(self.num_groups, 1, 1)
TypeError: 'module' object is not callable

Hi,
First of all, nice work!
I would like to obtain the source code to replicate the other experiments, such as the MNIST <-> SVHN. Could you make it available? Thanks.