-
-
I've got a question for the discriminator loss.
It seems when training using WGAN you can end up with increased image quality with increased loss.
I have plotted here -log D vs. generator iterat…
-
During the training of the generator, the gradients are calculated w.r.t D(G(z)) instead of -D(G(Z)).
-
Hi, I had test the tensorflow version and pytorch version of WGAN-GP and found that the pytorch version wasted more time. Is this normal or my test has something wrong? THANKS.
-
- [specNorm.pptx](https://github.com/ciplab/2019-winter-reading-project/files/4038364/specNorm.pptx)
- [wgan_20190923.pdf](https://github.com/ciplab/2019-winter-reading-project/files/4051954/wgan_201…
-
Hi,
In Table 1, the two FID scores for CelebA are completely identical to that of SVHN. From my own experiments, they are quite different. I would like to cross-check with you if this is an editin…
ghost updated
4 years ago
-
-
I download the LSUN bedroom dataset , the lmdb format is 54G, when I export it to image flie like png, It very huge maybe 500G , so do you use all Lsun bedroom images to train this model??
-
Like the title, in the WGAN-GP (improved WGAN) example, the paper states updating the critic n_critic=5 times for each generator update. This is not part of the current [example](https://github.com/te…
-
First, i have to thank for your sharing of this work, it does help me understanding the wgan-gp better, but here is a little question:
- what does the Mone in the loss_G.backward() mean?
l've ju…