search

AlexiaJM / RelativisticGAN

Code for replication of the paper "The relativistic discriminator: a key element missing from standard GAN"
719 stars 103 forks source link

Question #1

Closed taki0112 closed 6 years ago

taki0112 commented 6 years ago

Hi Very nice work. Thanks for sharing !

I have some questions

  1. when make the 256x256 images, you used PACGAN2 ! In your paper, you concat x1 & x2 x1 and x2 are same images ? or different image ?

  2. In RelativisticGAN, doesn't see mode collapse ?

  3. What is the performance of RLSGAN? I only see the RaLSGAN in your paper.

  4. Is batch_size not important for RelativisticGAN ?

Thank you

AlexiaJM commented 6 years ago
  1. Different images concatenated in the color channels. So the two images 3x256x256 now become 6x256x256. So you have to sample 2*batch_size images.
  2. There is some with 256x256 but not too much, FID would be high if there was a lot of mode collapse. See for yourself: https://imgur.com/a/pKI3PYa
  3. I did not try but it would be worth investigating. I only have a Geforce 1060 so I have to prioritize.
  4. I only used 32, but I assume not. For SGAN though larger batch size is a problem, it might worth comparing.
taki0112 commented 6 years ago

Thank you