List your company

Video Understanding on MMMU Video

55.91AccuracyHybrid-Frame Strategy
11.200422.807734.41546.0223Mar 18, 2026

Evaluation results

MethodLinks
Hybrid-Frame StrategyBackbone=Qwen2.5-VL-7B, Frames=64F2026.0355.91
Hybrid-Frame StrategyBackbone=Qwen2.5-VL-7B, Frames=32F2026.0355.12
Hybrid-Frame StrategyBackbone=Qwen2.5-VL-7B, Frames=16F2026.0351.78
Video-SFTBackbone=Qwen2.5-VL-7B, Frames=32F2026.0351.44
Video-SFTBackbone=Qwen2.5-VL-7B, Frames=64F2026.0351.44
Video-SFTBackbone=Qwen2.5-VL-7B, Frames=16F2026.0350.44
Video-SFTBackbone=Qwen2.5-VL-7B, Frames=8F2026.0347.56
Hybrid-Frame StrategyBackbone=Qwen2.5-VL-7B, Frames=8F2026.0347.22
Hybrid-Frame StrategyBackbone=LLaVA-1.5-7B, Note=Evaluated with a fixed 8-frame inference2026.0318.71
Video-SFTBackbone=LLaVA-1.5-7B, Frames=8F2026.0318.56
Video-SFTBackbone=LLaVA-1.5-7B, Frames=16F2026.0316.44
Video-SFTBackbone=LLaVA-1.5-7B, Frames=32F2026.0312.92