Text-to-Speech on French unseen languages
12.35WERGround Truth
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Ground TruthInference Mode=Zero-shot evaluation, Training Steps=300K steps2026.06 | 12.35 | — | 2.238 | 5.39 | |
| Pixel-TTSInference Mode=Zero-shot evaluation, Training Steps=300K steps2026.06 | 62.56 | 40 | 3.233 | 29.57 | |
| Text-TTSInference Mode=Zero-shot evaluation, Training Steps=300K steps2026.06 | 63.95 | 41.1 | 3.217 | 31.71 |