Apply Waseerstein GAN into SRGAN, a deep learning super resolution model

A PyTorch implementation of SRGAN specific for Anime Super Resolution based on "Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network". And another PyTorch WGAN-gp implementation of SRGAN referring to "Improved Training of Wasserstein GANs".

Applying Waseerstein GAN into SRGAN

No description available

Implemented a SRGAN model for Image Super Resolution. Reached a x4 upscaling factor with a PSNR of 29.23. Used pre-trained VGG-19 for the Discriminator and Wasserstein loss and developed a Weighted Dense Residual Network, among others improvements.