This repository contains PyTorch implementation of MD-GAN, along with training iPython notebook and trained models. Currently, this repository contains the training of data generated from a Gaussian mixture model (GMM). Two trained models included in this repository: the first one trained on data of a grid of 5 x 5 mixture of Gaussian and the second model trained on data of two mixture of Gaussian which are centered at -5 and 5.

The repository contains a training notebook for a Gaussian mixture model with 25 centers and a second notebook for Gaussian mixture model with two centers.

The trained model are located in this repository at the models folder.

Note: this code is not well documented yet, this will be done in the near future.

Example figures:

The left figure presents an example of real data and data generated using the MD-GAN D using the Gaussian mixture model with 25 and the right figure presents both the real data and generated data using MD-GAN while using the Gaussian mixture model with two centers.

If you find a bug or have a question, please create a GitHub issue.