gguogangg

just the same issue as caogang/wgan-gp#47 how do I understand the following code in https://github.com/KupynOrest/DeblurGAN/blob/2499a87f96f34f262ea294c7e4cd3fd7e90251f8/models/losses.py#L158 especially `dim=1`, it seems that the function calculate pixelwise gradient norm rather than the full...

it seems that the model will tend to make the token number larger when fix threshold (hard training step) because it cannot take L1 loss into account. How to solve...