Generative Adversarial Networks

Keras implementation of WassersteinGAN.

Sources:

• Wasserstein GAN

Requirements

python modules

• keras, theano or tensorflow backend
• h5py
• matplotlib
• opencv 3
• numpy
• tqdm
• parmap

Part 1. Processing the data

Follow these instructions.

Part 2. Running the code

Follow these instructions

Part 3. Example results

CelebA example results

For each image:

• The first 2 rows are generated images
• The last 2 rows are real images

MoG

Results on the unrolled GAN paper to dataset: