GANZoo

Environment Setup

Install Python 3.X and Pytorch according to your CUDA version Pip install ema-pytorch~=0.2.3 and tqdm

Train a GAN

For now, this code only implements training Wasserstein GAN with Gradient Penalty using MNIST, run this experiment by running the wgan.py file. You can call python wgan.py --params1 k1 --params2 k2 in terminal to customize your own parameters, where params1 or params2 are the parameter name and k1 or k2 are the parameter values.