Finetune for image-text retrieval task on COCO

salesforce / ALBEF

Code for ALBEF: a new vision-language pre-training method

BSD 3-Clause "New" or "Revised" License

1.53k stars 195 forks source link

Open yxoh opened 2 years ago

yxoh commented 2 years ago

Are the finetune settings exactly the same as in configs/Retrieval_coco.yaml and configs/Retrieval_flickr.yaml?

R@k became lower and lower when I set finetune with this.

LiJunnan1992 commented 2 years ago

Yes, you should be able to reproduce the paper's results by using the settings in the yaml files.