Video Understanding Reasoning on VideoMME Short
0.718AccuracySTD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| STDBackbone=Qwen2-VL-7B-Instruct2025.05 | 0.718 | 1.71 | |
| StreamingBackbone=Qwen2-VL-7B-Instruct2025.05 | 0.529 | 1.32 | |
| STDBackbone=LLaVA-OneVision-7B2025.05 | 0.518 | 1.82 | |
| StreamingBackbone=LLaVA-OneVision-7B2025.05 | 0.364 | 1.38 | |
| LayerSkipBackbone=LLaVA-OneVision-7B2025.05 | 0.056 | 0.33 | |
| LayerSkipBackbone=Qwen2-VL-7B-Instruct2025.05 | 0.037 | 0.59 |