UCMFH

When CLIP Meets Cross-modal Hashing Retrieval: A New Strong Baseline

Datasets

We release the three experimental datasets as follows:

• MIRFlickr, [Baidu Pan]
• NUS-WIDE, [Baidu Pan]
• MS COCO, [Baidu Pan]

Demo

Taking MIR Flickr as an example, our model can be trained and verified by the following command:

bash test-flickr.sh

Citation

If you use this code, please cite it:

@article{xia2023clip,
  title={When CLIP meets cross-modal hashing retrieval: A new strong baseline},
  author={Xia, Xinyu and Dong, Guohua and Li, Fengling and Zhu, Lei and Ying, Xiaomin},
  journal={Information Fusion},
  pages={101968},
  year={2023},
  publisher={Elsevier}
}