search

aitorzip / PyTorch-CycleGAN

A clean and readable Pytorch implementation of CycleGAN
https://arxiv.org/abs/1703.10593
GNU General Public License v3.0
1.19k stars 283 forks source link

why not fix D while updating G? #31

Closed XiaoyuShi97 closed 3 years ago

XiaoyuShi97 commented 3 years ago

Hi, thx for your nice work.

I have a question about training. When you train discriminator, you use detach which means avoid updating generator. But when you update generator, you do not fix discriminator. Is this a common strategy? Because I think in the original GAN paper, we fix discriminator too. https://github.com/aitorzip/PyTorch-CycleGAN/blob/67da8f9e2b69bd68763451803c7700aaccc92f18/train#L163

XiaoyuShi97 commented 3 years ago

Sorry. I see that these two nets use different optimizer.