Audio-Visual Captioning on DVD-Bench En
81.1AccuracyD-ORCA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| D-ORCANumber of Parameters=8B2026.02 | 81.1 | 16.6 | 57.1 | |
| AVoCaDONumber of Parameters=7B2026.02 | 72.9 | 17.9 | — | |
| Ours-SFT ModelNumber of Parameters=8B2026.02 | 71.2 | 29.8 | 31.8 | |
| Qwen3-Omni-InstructNumber of Parameters=30B-A3B2026.02 | 67.8 | 91.3 | — | |
| video-SALMONN 2+Number of Parameters=7B2026.02 | 66.6 | 94 | — | |
| ARC-Qwen-Video-NarratorNumber of Parameters=7B2026.02 | 66.4 | 65 | 23 | |
| Qwen2.5-OmniNumber of Parameters=7B2026.02 | 62.7 | 83.6 | — |