Multiple-choice Audio Question Answering on TUT 2017 (test)
78.4AccuracyVideoLLaMA2-AV
Evaluation Results
| Method | Links | |
|---|---|---|
| VideoLLaMA2-AVSize=7B, Training Hours=5k2024.06 | 78.4 | |
| VideoLLaMA2.1-AVSize=7B, Training Hours=5k2024.06 | 77.28 | |
| Qwen-AudioSize=7B, Training Hours=137k2024.06 | 64.9 |