Need README
#1
by massaki75 - opened
Hi, thanks for open-sourcing this work! In your paper, I see there are three variants of BMC-LongCLIP: one with 8K batch size, one with 16K batch size, and one trained with long captions. I wonder which one does the model released in this repository refers to?
We will be updating with more details soon! This one is the one with 8K and trained with the original captions but with extended context length. Here's how to use it: https://colab.research.google.com/drive/1TQ1SS_lXO-EalPT7h1CaZscWoYqdW0RE?usp=sharing
minwoosun changed discussion status to closed