Codes for RFSR: Improving ISR Diffusion Models via Reward Feedback Learning.
We release example Reward Feedback Learning training code in SeeSR.
To install the necessary dependencies, run the following commands:
pip install pyiqa
pip install image-reward
pip install git+https://github.com/fbcotter/pytorch_wavelets.gitbash SeeSR/train_seesr_kl_ema.sh
If our code helps your research or work, please consider citing our paper. The following are BibTeX references:
@article{sun2024rfsr,
title={RFSR: Improving ISR Diffusion Models via Reward Feedback Learning},
author={Sun, Xiaopeng and Lin, Qinwei and Gao, Yu and Zhong, Yujie and Feng, Chengjian and Li, Dengjie and Zhao, Zheng and Hu, Jie and Ma, Lin},
journal={arXiv preprint arXiv:2412.03268},
year={2024}
}
Visitor counter: