README.md

Generative Adverserial Networks

Goal will be to train two networks on MNIST data. One network will try to see if data comes from MNIST while the other network tries to generate images that look like MNIST

Generator

See the stand-alone genrator notebook

Discriminator

See the stand-alone discriminator notebook

GAN

See GAN notebook for the combined generator and discriminator models trained in altering manner

Linear GAN

The Linear GAN notebook shows a simple case of using a GAN to generate data that falls on a line.

Helper functions

The myHelper.py file contains helper code for getting the MNIST images.

Resources

• Original GAN Paper

• DCGAN Paper