search

google-research / composed_image_retrieval

Apache License 2.0
175 stars 18 forks source link

CLIP model #3

Closed LorenzoAgnolucci closed 1 year ago

LorenzoAgnolucci commented 1 year ago

Hi, thanks for publishing the code, I find your work very interesting!

I have one question: which CLIP model did you use to obtain the results reported in the paper? ViT-L/14 or ViT-L/14@336px?

Thanks in advance.

ksaito-ut commented 1 year ago

Thanks for your interest. We used ViT-L/14.

LorenzoAgnolucci commented 1 year ago

Thanks!