This is the official code of our work BlindDiff: Empowering Degradation Modelling in Diffusion Models for Blind Image Super-Resolution.
The pretrained model is here.
This repo is built on the basis of BasicSR and guided-diffusion, thanks for their open-sourcing!
- Python3
- pytorch>=1.7
Run the command:
pip install -r requirement.txt
and
python setup.py develop
- Download trainning dataset DIV2K and Flickr2K for the natural images and FFHQ for the face images.
- Configure
options/train.yml
for your training. - Run the command:
python basicsr/train.py -opt=options/train_setting.yml
- Configure
options/test.yml
for your training. The testing dataset used in the paper is here. - Run the command:
python basicsr/train.py -opt=options/test.yml