Speech to Sound generation on Speech-S
6.15WER (%)UniAudio 2.0 (Ours 3B)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| UniAudio 2.0 (Ours 3B)mode=Zero-shot, parameters=3B2026.02 | 6.15 | 0.11 | 2.96 | |
| UniAudio 2.0mode=Zero-shot2026.02 | 6.15 | 0.11 | 2.96 | |
| Ablation (Ours 1B)mode=Zero-shot, parameters=1B2026.02 | 6.79 | 0.04 | 2.82 |