Video-to-Audio generation on Multi-Caps VGGSound (test)
71.36Audio QualityMMAudio-S-16k + NAG
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MMAudio-S-16k + NAGBase Model=MMAudio-S-16k, Guidance=Negative Audio Guidance (NAG), Context=Composite audio2025.06 | 71.36 | 76 | 61.14 |