outputs_vlmeval/qwen2-vl-fromve/qwen2-vl-fromve_MMStar_acc.csv

"split","Overall","coarse perception","fine-grained perception","instance reasoning","logical reasoning","math","science & technology"
"none","0.452","0.596","0.392","0.612","0.444","0.404","0.264"