Implementation of MSCOCO retrieval metric

Open alex8937 opened this issue 4 years ago • 1 comments

Can the author confirm how the recall is implemented for both text to image and image to text given there are 5 captions per image?

Nov 17 '21 04:11 alex8937

Please check this: https://github.com/openai/CLIP/issues/115

May 05 '24 06:05 shyammarjit