Alpha Loss and Weight Loss

[stillbetter]

Hi Juewen, I want to know the specific implementation of these two loss functions? L1 regularization means l1_norm = torch.norm(alpha, p=1) or torch.nn.L1Loss(alpha,1)?