Spatio-Temporal Reasoning on ActivityNetQA
62.3AccuracyLLaVA-OneVision-72B
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaVA-OneVision-72BModel Scale=72B2024.08 | 62.3 | |
| GPT-4VVersion=V-Preview2024.08 | 57 | |
| LLaVA-OneVision-7BModel Scale=7B2024.08 | 56.6 | |
| LLaVA-OneVision-0.5BModel Scale=0.5B2024.08 | 50.5 |