Could you provide LLaVA-v1.5 7b/13b results on ChEF leaderboard?

Question

Could you provide LLaVA-v1.5 7b/13b results on ChEF leaderboard?

YuqiHUO opened this issue 10 months ago · 2 comments

Hi, I think ChEF is a great framework for evaluate large multi-modal LLM's. The result of LLaVA you provided in raw paper is LLaVA-1.0-mpt, which is a outdate version. I am curious to find out how does the latest LLaVA models perform in your framework, could you provide some result? Thanks!

Answer 1 · 2023-11-29T06:08:30.000Z

Thanks for your interest. We will deploy the latest LLaVA models, and provide the results on our leaderboard sooner or later.

Answer 2 · 2024-03-29T08:35:40.000Z

We provide a new leaderboard where LLaVA-v1.5 is evaluated. You can check our new work C $h^3$ Ef for more details link here.