Text-to-Speech on VCTK zero-shot
1.9WERGround Truth
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Ground TruthCodebook=N/S, Modality I/O=T,S | S2026.01 | 1.9 | — | — | — | — | |
| USLMCodebook=Dual, Modality I/O=T,S | S2026.01 | 6.5 | — | 11,611 | 0.89 | — | |
| AR-Omni (7B)Codebook=Single, Modality I/O=T,S,I | T,S,I2026.01 | 6.5 | 40 | 146 | 0.88 | — | |
| VALL-ECodebook=Dual, Modality I/O=T,S | S2026.01 | 7.9 | 600 | — | — | — | |
| AnyGPTCodebook=Dual, Modality I/O=T,S,I | T,S,I2026.01 | 8.5 | — | 6,602 | 2.92 | — | |
| MiOCodebook=Dual, Modality I/O=T,S,I | T,S,I2026.01 | 12 | 200 | 29,079 | 48.9 | — |