newbie_GAN Adversarial Autoencoder Tried to encode digits dataset into a swiss roll distribution. Result so far. Generative Adversarial Networks The principe works in general but to generate complicated forms like faces i need to build morecomplicated nets. Zero generation Faces generation (with weak net) Python 3.