AILab-CVC / SEED-Bench

(CVPR2024)A benchmark for evaluating Multimodal LLMs using multiple-choice questions.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Ask for original captions from which you generate questions

YangYangGirl opened this issue · comments

Congrats on the excellent work. Can you release the original captions from which you generate questions? It will be helpful to the community.

Thank you for your interest in our work. We find that the orignial captions from various vision foundation models can be noisy (that's why we use human annotation), so we may not release the original captions of this version.