search

LLaVA-VL / LLaVA-NeXT

Apache License 2.0
2.29k stars 154 forks source link

Eval results #106

Open liguopeng0923 opened 1 month ago

liguopeng0923 commented 1 month ago

Hi,

I evaluated your 7B model, but the results in new introduced benchmarks are not normal.

image image
FengLi-ust commented 1 month ago

Hi, thanks for this question. You are using the first version of the eval json. We have updated the evaluation json. Please download a new one from our website. LMK if you have further questions.

liguopeng0923 commented 1 month ago

Ok，I will update it. Thanks!

liguopeng0923 commented 1 month ago

Hi @FengLi-ust , I used the latest json file to evaluate the 0.5B model, but the results are still not normal.

image
FengLi-ust commented 1 month ago

How about the 7b results? Do they match the results in the table?

liguopeng0923 commented 1 month ago

It is abnormal too.

image
FengLi-ust commented 1 month ago

Hi, thanks for your feedback. We find out that the previous 7B model is not our best model and we opensource a wrong one. You can download our new 7B model (the same link as before) for test, which will match the results. Let me know if you have any questions.