Cross-modal retrieval on COCO

facebookresearch / vicreg

VICReg official code base

MIT License

516 stars 87 forks source link

Cross-modal retrieval on COCO #22

Open katerynaCh opened 1 year ago

katerynaCh commented 1 year ago

Hi! Thank you for your work! I have a question regarding cross-modal retrieval on COCO, I am struggling to reproduce the results reported in the paper. Could you provide more details on your training protocol? Which coefficients are you using for VICReg/what kind of expander architecture/are you doing any further downstream training or do you directly use the encoder embeddings obtained via ssl?

Thank you!

firasl commented 1 year ago

Hi, I am facing the same issue to reproduce the results of the paper. Did you find the right configuration?

MauritsBleeker commented 1 year ago

I am facing the same issue, too!