GitHub - johannnamr/Sinkhorn-GAN: Learning generative models with Sinkhorn Loss

johannnamr / Sinkhorn-GAN Public

forked from audeg/Sinkhorn-GAN

Notifications You must be signed in to change notification settings
Fork 0
Star 0

Learning generative models with Sinkhorn Loss

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
mke_tf.py		mke_tf.py
mke_tf.py_v2.py		mke_tf.py_v2.py
mmd.py		mmd.py
readme.txt		readme.txt
report.txt		report.txt
sinkhorn.py		sinkhorn.py
sinkhorn_gan_images.py		sinkhorn_gan_images.py
util.py		util.py

Repository files navigation

Two examples of learning with Sinkhorn divergences :

mke_tf.py : Tensor-flow implementation of learning a generative model for the MNIST dataset
-> code adapted from Jan Mentzen's implementation of a VAE (https://jmetzen.github.io/2015-11-27/vae.html)

sinkhorn_gan_images.py : pyTorch implementation of learning a generative model for the CIFAR dataset
(uses files sinkhorn.py, utils.py, mmd.py)
-> code adapted from MMD-GAN implementation (https://github.com/OctoberChang/MMD-GAN)