search

martinarjovsky / WassersteinGAN

BSD 3-Clause "New" or "Revised" License
3.2k stars 725 forks source link

Using batch normalization in the generator network leads to mode collapse #40

Closed yliu120 closed 7 years ago

yliu120 commented 7 years ago

I tried WGAN with Tensorflow and found that when batch normalization is not used in the generator (MLP network), then there is no mode collapse. However, when I added batch normalization between hidden layers, partial mode collapse occurred. I read your code and found that your MLP_G doesn't use any batch norm layers so I am wondering whether you have tried to add batch normalization in the generator only to see whether they will cause partial mode collapse.

Or, anyone already has some experience/ideas on this?

martinarjovsky commented 7 years ago

I haven't tried batch normalization with MLPs sadly. Sorry I can't be of further help.