VLMEvalKit
VLMEvalKit copied to clipboard
Ola模型自测与OpenCompass榜单差距很大
基于VLMEvalKit评测,裁判模型gpt-4-turbo
https://github.com/Ola-Omni/Ola/issues/14
@dongyh20 Could you please assist in resolving this question?
I will check this asap
We have reported the results on our own machine. The results have a slight difference with the official benchmark, but it is acceptable and it may caused by different machines or different envs. #https://github.com/Ola-Omni/Ola/issues/14