Zero-shot Text-to-Speech on Singing Voice
7.66WERVevo2
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Vevo2#Hz=12.5, Data (#hours)=101K, 7K, #Param=872M, PT?=✓2025.08 | 7.66 | 0.725 | 0 | 0 | |
| MaskGCT#Hz=50, Data (#hours)=101K, #Param=1048M, PT?=✗2025.08 | 11.71 | 0.753 | -1.74 | -0.97 | |
| Ground Truth2025.08 | 12.65 | 0.658 | 1.27 | 0.55 | |
| Vevo2-base#Hz=12.5, Data (#hours)=101K, 7K, #Param=872M, PT?=✗2025.08 | 15.78 | 0.708 | -0.28 | -0.09 | |
| F5-TTSData (#hours)=101K, #Param=336M, PT?=✗2025.08 | 16.12 | 0.597 | -1.89 | -1.29 | |
| CosyVoice 2#Hz=25, Data (#hours)=167K, #Param=581M, PT?=✓2025.08 | 16.18 | 0.659 | -1.68 | -1.12 | |
| Vevo2-base#Hz=12.5, Data (#hours)=101K, #Param=872M, PT?=✗2025.08 | 19.39 | 0.658 | -1.02 | -0.48 | |
| Vevo2-base#Hz=12.5, Data (#hours)=7K, #Param=872M, PT?=✗2025.08 | 25.38 | 0.692 | -0.69 | -0.36 |