Text-to-Speech on LEMAS-TTS zh (test)
1.38WERX-Voices1
Evaluation results
| Method | Links | ||
|---|---|---|---|
| X-Voices1Architecture=NAR, Parameters=0.3B, Supported Languages=30, Training Data=420K, Transcript-Free=true2026.05 | 1.38 | 0.816 | |
| X-Voices2Architecture=NAR, Parameters=0.4B, Supported Languages=30, Training Data=420K + 40K, Transcript-Free=false2026.05 | 1.87 | 0.817 | |
| LEMAS-TTSArchitecture=NAR, Parameters=0.3B, Supported Languages=10, Training Data=150K, Transcript-Free=true2026.05 | 2.17 | 0.788 |