A collection of generative methods implemented with TensorFlow (Deep Convolutional Generative Adversarial Networks (DCGAN), Variational Autoencoder (VAE) and DRAW: A Recurrent Neural Network For Image Generation).
Python
Latest commit 380b30c Jan 11, 2017 @ikostrikov committed on GitHub Merge pull request #11 from Banus/remove_pt_imports
remove remaining references to prettytensor

README.md

TF-VAE-GAN-DRAW

TensorFlow implementation of Deep Convolutional Generative Adversarial Networks, Variational Autoencoder (also Deep and Convolutional) and DRAW: A Recurrent Neural Network For Image Generation.

Run

VAE/GAN:

python main.py --working_directory /tmp/gan --model vae

DRAW:

python main-draw.py --working_directory /tmp/gan

Deep Convolutional Generative Adversarial Networks produce decent results after 10 epochs using default parameters.

TODO:

  • More complex data.
  • Add Adversarial Autoencoder
  • Replace current attention mechanism with Spatial Transformer Layer