bioinf-jku / TTUR

Two time-scale update rule for training GANs

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Score in Table 1

byeongkeun-kang opened this issue · comments

I have trained "GoodGenerator" (in the provided code) with TTUR (using the given default hyperparameters) for a couple of times. My trained model achieves about 29 FID score for the CIFAR-10 dataset. To achieve the score in Table 1, should I pick the best score among multiple training and multiple generations? Or, should I change the hyperparameters? Thank you.

Hi Byeongkeun, yes pick the best FID among all training runs and epochs. We can do that because we don't have a test set here. Greets Martin

@mhex Could you share your trained model that performs the best? I'm especially interested in WGAN_GP models. Thanks.

I'm sorry we can't send you the best model.