Mixed-Audio Generation on MECAT Speech + Music (SM0)
19.83FADVGGMusicGen
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MusicGen2026.05 | 19.83 | 57.28 | 2.38 | 13.8 | -4.99 | 99.99 | 1.65 | |
| TangoFlux2026.05 | 11.57 | 40.77 | 1.11 | 33.7 | 1.85 | 99.44 | 1.4 | |
| Qwen3-TTS2026.05 | 10.2 | 63.15 | 2.16 | 18.6 | -3.78 | 15.6 | 3.49 | |
| Expert-Pipeline2026.05 | 9.55 | 24.1 | 0.69 | 30.9 | 5.36 | 24.31 | 2.26 | |
| Dasheng AudioGen2026.05 | 1.7 | 6.69 | 0.33 | 32.7 | 9.8 | 21.96 | 2.72 |