You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi , I'm confused why use ''torch.sum(z**2)'' to calculate loss_ce. Besides, how can it be transformed from calculating loss_distr on Y x Z to calculating Loss_ce only on Z space?
The text was updated successfully, but these errors were encountered:
Please see Eq.(9) in the paper. z follows Gaussian distribution, so its log probability is ||z||^2. CE loss does not necessarily guarantee the distribution of (Y, Z) and is only a weaker surrogate loss for stable training (refer to #14).
Hi , I'm confused why use ''torch.sum(z**2)'' to calculate loss_ce. Besides, how can it be transformed from calculating loss_distr on Y x Z to calculating Loss_ce only on Z space?
The text was updated successfully, but these errors were encountered: