This repository holds the code to the "RenderGAN: Generating Realistic Labeled Data" paper.

You might be interested in:

• implementation of the RenderGAN
• code of the generator network
• discriminator
• architecture of the decoder network
• the Makefile to orchestrate the training process
• the hyperparameters for the training

We do not provide any information how to run our code. If you are interessted, fill an issue and let us know. This could motivate us enough to document the installation process and to upload the unlabeled training data and the real test data.