Audio Captioning on MECAT
60.11FENSE Content Long ScoreMiDashengLM
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| MiDashengLMModel Size=7B2025.08 | 60.11 | 61.38 | 50.69 | 53.78 | 66.17 | 51.06 | 32.4 | 59.5 | 51.38 | 57.53 | |
| Qwen2.5-OmniModel Size=7B2025.08 | 48.34 | 45.29 | 37.27 | 46.6 | 50.68 | 37.43 | 32.07 | 34.71 | 47.84 | 43.8 | |
| Kimi-Audio-InstructModel Size=7B2025.08 | 40.83 | 45.72 | 25.57 | 35.75 | 39.54 | 27.12 | 19.44 | 16.18 | 16.66 | 36.32 |