/MRefSR

LMR: A Large-Scale Multi-Reference Dataset for Reference-based Super-Resolution

Primary LanguagePythonApache License 2.0Apache-2.0

LMR: A Large-Scale Multi-Reference Dataset for Reference-based Super-Resolution

[Paper] [ArXiv] [Dataset Website]

Dependency

Environment

This code is based on PyTorch.

By using Anaconda, you can setup the environment simply by

conda env create -f environment.yml

Training

Train on the proposed LMR

./scripts/dist_train.sh 4 options/train/stage3_5ref_restoration_mse.yml

The results will be saved in ./experiments

Inference

Test on CUFED5 and the proposed LMR

python mmsr/test.py -opt test.yml

The results will be saved in ./result

Contact

Feel free to contact me if there is any question. (Lin Zhang, zhanglin2019@ia.ac.cn)