This respository uses a WGAN with gradient penalties to generate new pokemon. The model is still being worked on but some current results can be seen below:

To train the GAN run the notebook Generating-Pokemon.ipynb