mbzuai-oryx / LLaVA-pp

🔥🔥 LLaVA++: Extending LLaVA with Phi-3 and LLaMA-3 (LLaVA LLaMA-3, LLaVA Phi-3)
798 stars 58 forks source link

Clarification of ScienceQA score #2

Closed Isaachhh closed 5 months ago

Isaachhh commented 5 months ago

I notice that your model achieve ~80 on ScienceQA. I wonder whether you report total accuracy other than image accuracy, because LLaVA reports only image accuracy (66.8 for 7B and 71.6 for 13B). Thanks!

mmaaz60 commented 5 months ago

Hi @Isaachhh,

Thank you for pointing it out, and apologies for any confusion it may have caused. We have updated the main results and added some additional models as well. I am pasting the table below for your reference.

image

Let us know if you have any questions. Thank You

Isaachhh commented 5 months ago

Thank you!