Harry24k/adversarial-attacks-pytorch

[QUESTION] question about normalization

Opened this issue · 0 comments

kydxh commented

❔ Any questions

In the demo "demo/White-box Attack on ImageNet.ipynb", atk.set_normalization_used(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) is used. So after attack, the image will be normalized. In the code, imshow() was used to show the iamge. In this function, code "img = torchvision.utils.make_grid(img.cpu().data, normalize=True)" conduct another normalization. My question is why there need another normalize instead of inverse_normalize()?