Text-to-Speech (Text and Reference Speech input) on Emotion TTS ZH (Hard)
75.8ACC-IFlexiVoice
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| FlexiVoice2026.01 | 75.8 | 13.2 | 0.8 | 98.4 | |
| Groud-truth2026.01 | 61.6 | 4.4 | 1 | — | |
| CosyVoice22026.01 | 47.8 | 15.3 | 0.79 | 100 | |
| FlexiVoice-Base2026.01 | 22.4 | 38 | 0.74 | 99.2 | |
| VoxInstruct2026.01 | 18.7 | 23.2 | 0.73 | 59.8 |