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

OpenGVLab / LAMM

[NeurIPS 2023 Datasets and Benchmarks Track] LAMM: Multi-Modal Large Language Models and Applications as AI Agents

https://openlamm.github.io/

297 stars 16 forks source link

Could you provide LLaVA-v1.5 7b/13b results on ChEF leaderboard? #53

Closed YuqiHUO closed 6 months ago

YuqiHUO commented 10 months ago

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!

Coach257 commented 10 months ago

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

Coach257 commented 6 months ago

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.