[Result] Can you provide some results on benchmarks?

[Chenfeng1271]

Hi,

Thank you for the excellent work, when I evaluate the model on different benchmarks, I find some results are unreasonable, so can u provide your results to help me figure out the correct config? Thank you.

[Luodian]

Thanks for this suggestion. Providing a comprehensive table is challenging, but we are trying to make it.

Can you first point out which result is unreasonable?

We can help you to figure out.

[Chenfeng1271]

Hi, I run llava1.5-7b on ICON-QA but get the result is 0

[Chenfeng1271]

I also run llava1.5-7b on MMVet and get 30.367, but the official result is 43.9

[Chenfeng1271]

The evaluation on hallusion_bench_image has a bug