• [2025.04.21] ✅ Upload our model on Hugging Face 🤗.
• [2025.04.15] 🎉 Our paper is accepted to CVPR 2025 Workshop!
• [2025.03.26] 🏆 Our team won 1st place in the NTIRE 2025 Efficient SR Challenge. Challenge report is here.
• [2025.03.21] ✅ Release our code here.

The evaluation environments adopted by us is recorded in the requirements.txt. After you built your own basic Python (Python = 3.9 in our setting) setup via either virtual environment or anaconda, please try to keep similar to it via:

• Step1: install Pytorch first: pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu117

• Step2: install other libs via: pip install -r requirements.txt

or take it as a reference based on your original environments.

1. Run the run.sh

   CUDA_VISIBLE_DEVICES=0 python test_demo.py --data_dir [path to your data dir] --save_dir [path to your save dir] --model_id 23

   • Be sure the change the directories --data_dir and --save_dir.

If our work is useful to you, please use the following BibTeX for citation.

@inproceedings{Chai2025DistillationSupervisedCL,
  title={Distillation-Supervised Convolutional Low-Rank Adaptation for Efficient Image Super-Resolution},
  author={Xinning Chai and Yao Zhang and Yuxuan Zhang and Zhengxue Cheng and Yingsheng Qin and Yucai Yang and Li Song},
  year={2025},
  url={https://api.semanticscholar.org/CorpusID:277787382}
}

This code repository is release under MIT License.