Motion Understanding on AV-Hand
72.7AccuracyGemini-2.5-Flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-2.5-FlashMode=Zero-shot2025.12 | 72.7 | |
| NVILA-Video-8BMode=Fine-tuned, Data=FoundationMotion2025.12 | 61.4 | |
| NVILA-Video-15BMode=Fine-tuned, Data=FoundationMotion2025.12 | 58.7 | |
| NVILA-Video-15BMode=Zero-shot2025.12 | 58.1 | |
| Qwen-2.5-VL-72BMode=Zero-shot2025.12 | 56.5 | |
| NVILA-Video-8BMode=Fine-tuned, Data=PLM 467k2025.12 | 56 | |
| NVILA-Video-15BMode=Fine-tuned, Data=PLM 467k2025.12 | 55.6 | |
| NVILA-Video-8BMode=Zero-shot2025.12 | 54.6 | |
| Qwen-2.5-VL-7BMode=Fine-tuned, Data=FoundationMotion2025.12 | 52.8 | |
| Qwen-2.5-VL-7BMode=Zero-shot2025.12 | 47.2 |