pytorch-domain-adaptation icon indicating copy to clipboard operation
pytorch-domain-adaptation copied to clipboard

Published 20 hours ago • verified publisher icon jvanvugt

questions about loss？

Open Dr-Zhou opened this issue 5 years ago • 2 comments

Dr-Zhou avatar Sep 24 '19 06:09 Dr-Zhou

why the loss don't include critic loss?The paper final loss =classifier loss + wasserstain loss +gradient penalty loss.

Dr-Zhou avatar Sep 24 '19 07:09 Dr-Zhou

I have the same confusion. The model is initiated by the pre-trained model on the source domain, which is different from the original paper.

weiyikang avatar Sep 21 '20 12:09 weiyikang