Breakdown of the score on benchmarks

[mojtaba-komeili]

I was wondering if you have the breakdown of the scores on some of the performance benchmarks. This is interesting to see for some of the benchmarks that have sub-tasks. More specifically I am interested to know that on MVBench and Perception Test as you have on table 5 in LLaVA-OneVision paper.