---
pipeline_tag: image-to-image
tags:
- blind-image-inpainting
- inpainting
- image-to-image
---
# CAML Pre-trained Checkpoints (Unofficial Mirror)

> **⚠️ Disclaimer:** I am **not** affiliated with the authors of this work. This repository is an unofficial mirror of the pre-trained model weights to improve accessibility for researchers outside China who may have difficulty accessing Baidu Cloud.

## Original Work

**Paper:** [Context-Aware Mutual Learning for Blind Image Inpainting and Beyond](https://github.com/zhenglab/CAML)

**Authors:** Haoru Zhao, Zhaorui Gu, Bing Zheng, Haiyong Zheng  
*Ocean University of China*

**Official Repository:** https://github.com/zhenglab/CAML

**Original Checkpoint Source:** Baidu Cloud (linked in official repo), retrieved on January 14, 2026 via [this link](https://pan.baidu.com/s/1-VzOl35c7NYaUprAwRXoQw?pwd=7qmx) (extraction code: `7qmx`)

## Citation

If you use these weights, please cite the original authors' work:

```bibtex
@article{zhao2023caml,
  title={Context-Aware Mutual Learning for Blind Image Inpainting and Beyond},
  author={Zhao, Haoru and Gu, Zhaorui and Zheng, Bing and Zheng, Haiyong},
  journal={Neural Networks},
  year={2023}
}
```

## License

Please refer to the [official repository](https://github.com/zhenglab/CAML) for licensing information. This mirror is provided solely for research accessibility and all rights remain with the original authors.

## Issues

For issues related to the model architecture or training, please use the [official GitHub repository](https://github.com/zhenglab/CAML/issues).