Zero-shot Text-to-Speech on CV3-Eval Multilingual Voice Cloning
4.87English CER/WERCosyEdit2
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CosyEdit22026.05 | 4.87 | 3.52 | 6.16 | 5.14 | |
| CosyVoice22026.05 | 5.45 | 3.77 | 7.76 | 6.89 | |
| CosyEdit2026.05 | 7.22 | 5.29 | 17.31 | 9.25 | |
| SSR-Speech2026.05 | 14.08 | — | — | — | |
| VoiceCraft-X2026.05 | 19.17 | 9.74 | 23 | 43.15 |