JackBurdick / GAN_face_generation

generate novel faces and numerical digits using generative adversarial networks in tensorflow

Generative Adversarial Network (GAN) implemented in TensorFlow

GAN is created and explored on two data sets (MNIST and a celebrity photo data set) using TensorFlow.

Examples

Novel handwritten digits are produced after training on the MNIST data set.

Novel images of faces are generated after training the GAN on a celebrity photo data set.

Project Information

The included ./dlnd_face_generation.ipynb provides a commented walk though of the network generation and evaluation.

References:

• tensor flow documentation

Note: Project created by Udacity as part of DLND