Video Question Answering on VIDAL-QA LLaVA-Hound in-domain (test)
68.5AccuracyMM1.5-Video
Evaluation Results
| Method | Links | |
|---|---|---|
| MM1.5-VideoModel Scale=7B, Protocol=SFT2024.09 | 68.5 | |
| MM1.5-VideoModel Scale=3B, Protocol=SFT2024.09 | 63.4 | |
| MM1.5-VideoModel Scale=1B, Protocol=SFT2024.09 | 60.6 | |
| LLAVA-HOUND-SFTModel Scale=7B2024.09 | 56.3 | |
| MM1.5-VideoModel Scale=7B, Protocol=Training-free2024.09 | 48.7 | |
| MM1.5-VideoModel Scale=3B, Protocol=Training-free2024.09 | 45.4 | |
| MM1.5-VideoModel Scale=1B, Protocol=Training-free2024.09 | 42.6 | |
| Video-LLaVAModel Scale=7B2024.09 | 34.3 | |
| Chat-UniViModel Scale=7B2024.09 | 31.4 | |
| LLaMA-VIDModel Scale=7B2024.09 | 30.6 | |
| Video-ChatGPTModel Scale=7B2024.09 | 29.4 |