• You can find the datasets from Google Drive link.
• Full code has been released. You can train it or evaluate with our provided pretrained model weights from Google Drive link

Note: The project is finally open-sourced! Sorry it took so long, and the delay was about half a year, which was not as expected. Since I am not a student now, I got busy with work and honestly forgot about it for a while. I didn’t have the chance to maintain or clean up the code, but I still hope it proves useful to you.

The CUDA driver must be higher than 11.6

• Ubuntu 20.04
• CUDA 11.8
• Python 3.9
• PyTorch 2.1.1 + cu118

1: First install Pytorch (2.1.1+cu118) as following command:

pip install torch==2.1.1 torchvision==0.16.1 torchaudio==2.1.1 --index-url https://download.pytorch.org/whl/cu118

2: Install other requiremnets:

pip install -r requirements.txt

Please cite us if this work is helpful to you.

@inproceedings{li2024ucip,
  title={UCIP: A universal framework for compressed image super-resolution using dynamic prompt},
  author={Li, Xin and Li, Bingchen and Jin, Yeying and Lan, Cuiling and Zhu, Hanxin and Ren, Yulin and Chen, Zhibo},
  booktitle={European Conference on Computer Vision},
  pages={107--125},
  year={2024},
  organization={Springer}
}

The basic code is partially from the below repos.

• ActiveMLP