search

aryopg / Feature_Matching_GAN

0 stars 0 forks source link

About Custom JSD Loss #1

Open zjzzzz opened 6 years ago

zjzzzz commented 6 years ago

Excuse me.Can you tell me something about the Custom JSD Loss? Which paper adress the formula or can you list the formula?I'm a little confused.And thanks for sharing the code with us~

看你头像好像是中国人，如果可以的话能否加个微信或者qq一起探讨一下~。

aryopg commented 6 years ago

Hi @zjzzzz . Thanks for your interest of this repo. The custom JSD Loss was taken from a paper for NLP. This is still a work in progress, I'm trying to validate the effectivity of this loss using the MNIST data, but the end goal is not disclosable yet (just got rejected from a conference). Will share more details ASAP :)