haotian-liu / LLaVA

[NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.
https://llava.hliu.cc
Apache License 2.0
17.84k stars 1.93k forks source link

[Discussion] why are the results on post inconsistent with opencompass? #1556

Open geweihgg opened 2 weeks ago

geweihgg commented 2 weeks ago

Discussion

post:https://llava-vl.github.io/blog/2024-01-30-llava-next/ MMVet = 48.4 image

opencompass:https://rank.opencompass.org.cn/leaderboard-multimodal/?m=REALTIME MMVet = 44.9 image

Is there any difference such as model version, test strategy?