ELIP: Efficient Language-Image Pre-training with Fewer Vision Tokens

Code contribution: Yangyang Guo 60% and Haoyu Zhang 40%.

Detailed configuration of each model can be found on each model directory.

Pre-trained Checkpoints

If you have a Baidu Drive account, check this link.

Citation:

If you found this repo helpful, please consider cite the following paper 👍 :

@inproceedings{elip,
  author    = {Guo, Yangyang and Zhang, Haoyu and Nie, Liqiang and Wong, Yongkang and Kankanhalli, Mohan},
  title     = {ELIP: Efficient Language-Image Pre-training with Fewer Vision Tokens},
  booktitle = {arXiv preprint arXiv:2309.16738},
  year      = {2023}}