rosinality/swapping-autoencoder-pytorch

Generator output

tom99763 opened this issue · 0 comments

I just want to ask why you don't use tanh or sigmoid for output activation