Hi, @beichenzbc
Thank you for your amazing work on Long-CLIP, which has greatly inspired me to improve CLIP's max token limitation and enhance image-retrieval performance.
I have two questions to discuss with you:
According to the paper, long and short captions are needed for training Long-CLIP. However, I noticed that ShareGPT-4V only provides long captions. How can we generate or obtain short captions?
Do you have any plans to release the Long-CLIP training code?
Hi, @beichenzbc Thank you for your amazing work on Long-CLIP, which has greatly inspired me to improve CLIP's max token limitation and enhance image-retrieval performance.
I have two questions to discuss with you:
Thank you.