A PyTorch implementation of the OT-GAN paper. OT-GAN is a variant of generative models that uses Sinkhorn Divergence to minimize the distance between the generator and data distribution.
cd ot-gan
pip install -r requirements.txt- Salimans, Tim, et al. Improving GANs Using Optimal Transport. arXiv, 14 mars 2018. arXiv.org, https://doi.org/10.48550/arXiv.1803.05573.