Video Question Answering on VideoQA Average
3.51ScoreLLaVolta
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LLaVolta#Stages=Three, Scheme=compr. deeper, #Tokens=84776, CR=174%, TFLOPs=17.29, Train-time=37.1h2024.06 | 3.51 | 0.58 | |
| LLaVolta#Stages=Three, Scheme=compr. wider, #Tokens=83256, CR=177%, TFLOPs=16.86, Train-time=37.0h2024.06 | 3.51 | 0.58 | |
| LLaVolta#Stages=Two, Scheme=compression, #Tokens=80496, CR=183%, TFLOPs=17.73, Train-time=37.1h2024.06 | 3.5 | 0.579 | |
| LLaVolta#Stages=Four, Scheme=wider then deeper, #Tokens=88704, CR=166%, TFLOPs=18.32, Train-time=37.2h2024.06 | 3.5 | 0.581 | |
| LLaVolta#Stages=Four, Scheme=deeper then wider, #Tokens=86904, CR=170%, TFLOPs=18.64, Train-time=37.1h2024.06 | 3.5 | 0.582 | |
| VideoLLaVA#Stages=Single, Scheme=no compression, #Tokens=147456, TFLOPs=29.88, Train-time=40.7h2024.06 | 3.48 | 0.578 | |
| VanillaRetained Tokens=2,048, Token Reduction Ratio=100%2026.04 | 2.55 | 46.1 | |
| LearnPrunerRetained Tokens=456, Token Reduction Ratio=77.8%2026.04 | 2.51 | 45.5 | |
| LearnPrunerRetained Tokens=228, Token Reduction Ratio=88.9%2026.04 | 2.45 | 43.9 | |
| FastVRetained Tokens=456, Token Reduction Ratio=77.8%2026.04 | 2.42 | 43.2 | |
| FastVRetained Tokens=228, Token Reduction Ratio=88.9%2026.04 | 2.21 | 39.3 |