General Video Understanding on TempCompass (test)
69.7AccuracyNVILA-8B
Evaluation results
| Method | Links | |
|---|---|---|
| NVILA-8BBackbone=8B2026.05 | 69.7 | |
| LLaVA-Video-7B + ST-GridPoolInput frames=64, Pooling Strategy=ST-GridPool2026.05 | 66.1 | |
| LLaVA-Video-7BInput frames=642026.05 | 65.4 | |
| Apollo-7BBackbone=7B2026.05 | 64.9 | |
| LLaVA-OneVision-7B + ST-GridPoolInput frames=32, Pooling Strategy=ST-GridPool2026.05 | 64.4 | |
| LLaVA-OneVision-7BInput frames=322026.05 | 64.2 | |
| VideoLLaMA2.1-7BBackbone=7B2026.05 | 56.8 |