This code implements a vanilla GAN, as described in the original paper by Goodfellow et al.. It implements both the generator and the discriminator as fully connected feedforward networks, without using any deep learning libraries like TensorFlow or Theano. Both implementations are heavily derived from Denny Britz's excellent tutorial.
This source code was originally developed as the final project for the course CS674: Mathematical Topics In AI & Optimization (Fall 2016) at Rutgers.