-
I think our current model with `TensorGraph` can't nicely support GANs. In order to train GANs, you need to train both a discriminator `D` and a generator `G`. The training of these two models is thre…
-
When I implement this paper ["LR-GAN: Layered Recursive Generative Adversarial Networks for Image Generation"](https://arxiv.org/pdf/1703.01560.pdf) . I meet the error:
```
NotImplementedError …
-
source code:
https://github.com/bioinf-jku/TTUR/blob/ca9b6572d08f81d0725de8558400fb17585266d3/WGAN_GP/gan_64x64_FID.py#L45-L49
-
Default kernel weights initialization of convolution layer. I use the function conv2d, but I can't find the initial weights of the convolution kernel , or how initialize the weights of convolution ke…
-
In the dcgan example, while training the discriminator, why is backward called twice ? First its called on the real images, then the fake images.
Instead, shouldn't doing something like:
`totalErr…
-
for the wgan-gp?
Thanks!
ypxie updated
7 years ago
-
When I try to run dcgan from pytorch/examples, I selected the cifar 10 dataset. It took very long time to load cifar 10 into gpu memory while cifar 10 is relatively small dataset compared with dataset…
ghost updated
7 years ago
-
Thanks for the PyTorch implementation of DiscoGAN.
I am having troubles running it with multiple GPUs though.
I think the first problem is in the `config.py` file, where it casts the `num_gpu` argum…
-
Is their any examples with the usage of Multiprocessing? It should be faster to use one thread or process to generate batch and manipulate the data.
-
First, thanks for creating such a focused, bounded, well implemented, library. It is fantastic to work with. Highway networks might be a useful feature for `torch.nn`. They are needed to implement pap…