Text-to-Speech Synthesis on Internal (test)
4.16MOSRecording
Evaluation results
| Method | Links | |
|---|---|---|
| Recording2026.06 | 4.16 | |
| dynamic prosody predictionTraining dataset=50k hours, alpha=0.52026.06 | 3.99 | |
| CoTTraining dataset=50k hours2026.06 | 3.98 | |
| CosyVoice(50k)Training dataset=50k hours2026.06 | 3.97 |