Poor results?

[gigadeplex]

After training it on 110k iterations (paper trained it on 100k) this is the result. It seems very different from the paper, have I done something wrong? I must add that I trained it on normal celeba not hq.

[elvisyjlin]

Actually, we're still working on this repo. But I'm busy with other things so it could take a few weeks. We assure that you can get the same result as those on the paper from https://github.com/willylulu/RelGAN, which is our original codebase. This PyTorch version is translated from the Keras one.

[HelenMao]

Hi, I want to know will they have much difference in the results of the celebA dataset between that of the celebA-HQ dataset?

[KelestZ]

I met the same issue using normal CelebA and seemingly it doesn't work at all...