# PW-FNet

**Repository Path**: atari/pw-fnet

## Basic Information

- **Project Name**: PW-FNet
- **Description**: 同步 https://github.com/deng-ai-lab/PW-FNet
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-01-09
- **Last Updated**: 2026-01-09

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

## [Global Modeling Matters: A Fast, Lightweight and Effective Baseline for Efficient Image Restoration] (Under Review)
 Official implementation.
 
#### News
Since our paper is currently under review, we are only open-sourcing the rain removal-related code at this stage. Once the paper is accepted, we will open-source all the code. If you have any other academic-related requests or discussions, please feel free to raise them in the issue section or contact me via email. 

**We sincerely hope that if you find our work useful, you will kindly star our project and cite our paper appropriately! Thank you very much!**

- **Aug 13, 2025:** Our paper is available at [https://arxiv.org/pdf/2507.13663](https://arxiv.org/pdf/2507.13663).

<hr />

## Prepare Datasets

**Deraining Datasets:** <a href="https://pan.baidu.com/s/1OCaSuwQRgkpm3kU55Hef0Q?pwd=sakm">Rain200L/Rain200H</a> <a href="https://pan.baidu.com/s/1zknuXnv5MR6yhS0I8u2uZg?pwd=h8vv">DDN-Data</a> <a href="https://pan.baidu.com/s/1RsDzu_lS-l2_FxdUnBarOQ?pwd=vxgq">DID-Data Train</a> <a href="https://pan.baidu.com/s/1sJUb6iwxAWOcoBBLn0iuUA?pwd=6bpj">DID-Data Test</a> <a href="https://pan.baidu.com/s/1ESuLn9a9G0QRD22MLxDPxw?pwd=bnqh">SPA-Data</a> <a href="https://pan.baidu.com/s/1zztcAht2aQje4hHcDnGynQ?pwd=2y7d">Raindrop</a>

**Motion Deblur Datasets:** <a href="https://pan.baidu.com/s/1ANXeNVc-sySlqgpOSaq9vQ?pwd=c5jc">Motion Blur(GoPro)</a>

**Super-resolution Datasets:** <a href="https://pan.baidu.com/s/10OaDk8DttGzHO6Bh9nnPFw?pwd=68gw">DIV2K Set5 Set14 B100 Urban100 Manga109</a>

**More Datasets:** Todo

## Pretrained model
**Deraining models:** <a href="https://pan.baidu.com/s/1XFoItAezRZEHF2-c_jn-jA?pwd=ex3b">Rain200L</a> <a href="https://pan.baidu.com/s/1cat77YBYgW48TkLIH86UMQ?pwd=vx5j">Rain200H</a> <a href="https://pan.baidu.com/s/1dYpfoSUOMzJBI6J4pY4_4w?pwd=tsrs">DID</a> <a href="https://pan.baidu.com/s/1BxucIQVSAxzBVSqPxAplLQ?pwd=evwt">DDN</a> <a href="https://pan.baidu.com/s/1lrsQ4lazPh4lC3fjxEV_eQ?pwd=kr7s">SPA-Data</a>

**More models:** Todo

## Visual Results
<table>
<thead>
  <tr>
    <th>Deraining Dataset</th>
    <th>Rain200L</th>
    <th>Rain200H</th>
    <th>DID</th>
    <th>DDN</th>
    <th>SPA-Data</th>
    <th>Raindrop</th>
    <th>4K-Rain13k</th>
  </tr>
</thead>
<tbody>
  <tr>
    <td>Baidu NetDisk</td>
    <td> <a href="https://pan.baidu.com/s/1tdvEezksw3lpUsHgzWXmfA?pwd=jkm8">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/1YtCwcFPfyC0tOJULvwW3_A?pwd=u8se">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/14QYesghoCOLN29nJPA92eA?pwd=84mm">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/1dbOT2-0WKl-DqmAB4TcSOw?pwd=j2y3">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/1nPrVMIQw_htFm4eGXpwiww?pwd=wdpu">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/1EQH7Xvpb3Wd2LbKiovwefQ?pwd=dvg2">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/1UNfsIIVGjPtqTfHM8VkTrg?pwd=cmyv">Download</a>  </td>
  </tr>
</tbody>
</table>

<table>
<thead>
  <tr>
    <th>Super-resolution Dataset</th>
    <th>X2</th>
    <th>X3</th>
    <th>X4</th>
  </tr>
</thead>
<tbody>
  <tr>
    <td>Baidu NetDisk</td>
    <td> <a href="https://pan.baidu.com/s/1ox-r4pIuxly4THx46oUt0Q?pwd=v39n">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/109Ik730MS-u705NRauhwyA?pwd=2b2h">Download</a>  </td>
    <td> <a href="https://pan.baidu.com/s/1xmxFYV-kj2TuHWU7K_tvzg?pwd=rtj3">Download</a>  </td>
  </tr>
</tbody>
</table>

<table>
<thead>
  <tr>
    <th>Motion Deblurring Dataset</th>
    <th>GoPro</th>
  </tr>
</thead>
<tbody>
  <tr>
    <td>Baidu NetDisk</td>
    <td> <a href="https://pan.baidu.com/s/1CzTKvR8PTLZD4SZY_DJ7Hg?pwd=cayf">Download</a>  </td>
  </tr>
</tbody>
</table>


## References
If you find our work useful, please cite our paper:

    @article{jiang2025global,
      title={Global Modeling Matters: A Fast, Lightweight and Effective Baseline for Efficient Image Restoration},
      author={Jiang, Xingyu and Gao, Ning and Dou, Hongkun and Zhang, Xiuhui and Zhong, Xiaoqing and Deng, Yue and Li, Hongjue},
      journal={arXiv preprint arXiv:2507.13663},
      year={2025}
    }

## Contact

If your submitted issue has not been noticed or there are further questions, please contact jxy33zrhd@buaa.edu.cn.