Tensorflow Implementation on "The Cramer Distance as a Solution to Biased Wasserstein Gradients" (https://arxiv.org/pdf/1705.10743.pdf)
Another recent member in the GAN family!
Seems to be an improvement over Improved WGAN (Improved^2 GAN)?
To run an example for MNIST generation:
python cramer-gan.py
This is a modification of Improved WGAN, where the losses are changed according to the following:
Tensorflow Implementation on "The Cramer Distance as a Solution to Biased Wasserstein Gradients" (https://arxiv.org/pdf/1705.10743.pdf)