AST

KO | EN

gitlite — search

#typescript #ai-agents #ai #dsh-plugin #deepseek-harness #open-source #cli #claude-code #codex #developer-tools #react #windows

★ 179 Open GitHub ↗

Adapt or Perish: Adaptive Sparse Transformer with Attentive Feature Refinement for Image Restoration

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Adapt or Perish: Adaptive Sparse Transformer with Attentive Feature Refinement for Image Restoration (CVPR 2024)

[Shihao Zhou](https://joshyzhou.github.io/), [Duosheng Chen](https://github.com/Calvin11311), [Jinshan Pan](https://jspan.github.io/), [Jinglei Shi](https://jingleishi.github.io/), and [Jufeng Yang](https://cv.nankai.edu.cn/)

#### News
- **Jul 16, 2025:** [Hugging Face Demo](https://huggingface.co/spaces/suncongcong/AST-Demo) is available now, thanks contribution of [congcong](https://github.com/naninasm)
- **Feb 27, 2024:** AST has been accepted to CVPR 2024 :tada:

<hr />

## Package dependencies
The project is built with PyTorch 1.9.0, Python3.7, CUDA11.1. For package dependencies, you can install them by:
```bash
pip install -r requirements.txt
```
## Training
### Derain
To train AST on SPAD, you can run:
```sh
sh script/train_derain.sh
```
### Dehaze
To train AST on Densehaze, you can run:
```sh
sh script/train_dehaze.sh
```
### Raindrop
To train AST on AGAN, you can run:
```sh
sh script/train_raindrop.sh
```

## Evaluation
To evaluate AST, you can run:

```sh
sh script/test.sh
```
For evaluate on each dataset, you should uncomment corresponding line.

## Results
Experiments are performed for different image processing tasks including, rain streak removal, raindrop removal, and haze removal. 
Here is a summary table containing hyperlinks for easy navigation:

<table>
 <tr>
 <th align="left">Benchmark</th>
 <th align="center">Pretrained model</th>
 <th align="center">Visual Results</th>
 </tr>
 <tr>
 <td align="left">SPAD</td>
 <td align="center"><a href="https://pan.baidu.com/s/1uqsXeztA55ny8vnQHfxmlA?pwd=h68m">(code:h68m)</a></td>
 <td align="center"><a href="https://pan.baidu.com/s/1KtAWESp3mzfrV4qy7BLaWg?pwd=wqdg">(code:wqdg)</a></td>
 </tr>
 <tr>
 <td align="left">AGAN</td>
 <td align="center"><a href="https://pan.baidu.com/s/1obhEpvWLV9FLF47FMcNX7g?pwd=astt">(code:astt)</a></td>
 <td align="center"><a href="https://pan.baidu.com/s/1VewSQl6c7uWHpfCDVQeW4g?pwd=astt">(code:astt)</a></td>

</tr>
 <tr>
 <td align="left">Dense-Haze</td>
 <td align="center"><a href="https://pan.baidu.com/s/1PeSaf-ZRwnMBjlNHGEkvzg?pwd=astt">(code:astt)</a></td>
 <td align="center"><a href="https://pan.baidu.com/s/18JK8hQFLzSuiVZyianmJQw?pwd=astt">(code:astt)</a></td>
 </tr>

</table>

## Citation
If you find this project useful, please consider citing:

@inproceedings{zhou2024AST,
      title={Adapt or Perish: Adaptive Sparse Transformer with Attentive Feature Refinement for Image Restoration},
      author={Zhou, Shihao and Chen, Duosheng and Pan, Jinshan and Shi, Jinglei and Yang, Jufeng},
      booktitle={CVPR},
      year={2024}
    }

## Acknowledgement

This code borrows heavily from [Uformer](https://github.com/ZhendongWang6/Uformer).

AST

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README