This repository is developed by @penguin1214 and @Paper99.

Our super-resolution framework is built on BasicSR / EDSR-Pytorch / RCAN and tested on Ubuntu 14.04/16.04 environment (Python3.6, PyTorch 0.4.0, CUDA9.0/8.0, cuDNN7/5.1) with NVIDIA GPUs

• Python 3.6
• Pytorch 0.4.0

• TensorFlow (for better logging)

You should download some SR datasets firstly, for example:

• training datasets: DIV2K (800 images for train, 100 images for validation/test). It can be downloaded from official site; (If you just want to test our SR framework, you can download debug dataset from here)
• testing datasets: Set5, Set14, Urban100, B100, Manga109; (The download link is provided by @yulunzhang)