Zhenliang He

@jiaoxijie I am sorry that I am too busy these days to answer your questions. Your questions are not easy to answer because many situations cause your problems. First, what...

@Curiosit Did you put the data into the correct place? Or did you totally follow the REAME.md steps?

@supritbansod Try to decrease cycle_loss_weight（or increase the adversarial loss weight, hard coded as 1） or train it with a sufficiently long time. But I am not sure whether it will...

@NabHamzah Can you show me some cases？

@sangjun04 Maybe the droplets are caused by the architecture? I am not sure. You can refer to StyleGAN v2 paper which solves the droplet problem caused by instance normalization.

@OriYitzhaki Try to train D 5 times since WGAN need more D training. I.e., change this line https://github.com/LynnHo/CycleGAN-Tensorflow-2/blob/088263857d2fb9b8f06caa281b62bcad43eddadc/train.py#L149 to be ```python for _ in range(5): D_loss_dict = train_D(A, B, A2B,...

@OriYitzhaki I will try to figure out the problem. It seems that there are some bugs in WGAN-GP with TF2.

@dzlab Thanks for the suggestion, I will add it in the spare time.

@RayGuo-C Because G and D training are separated，they don't need the same noise input.

@brunolemos The desktop software turns the login to the browser but nothing happens in the software after I finish login. Does anybody have the same situation?