$ cd Awesome-GANs
$ python3 -m awesome_gans.wgan [parameters...]
$ cd Awesome-GANs
$ python3 -m awesome_gans.wgan --width 32 --height 32 --n_channels 3 --dataset 'cifar10'
$ cd Awesome-GANs
$ python3 -m awesome_gans.wgan --width 28 --height 28 --n_channels 1 --dataset 'mnist'
- Same with the
WGAN
paper.
Difference | Paper | Mine |
---|---|---|
Networks | ResNet like |
simple 4 conv nets |
Name | Epoch 1 | Epoch 20 | Epoch 1 ~ 35 |
---|---|---|---|
WGAN |