Mandarin
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Tone Geometry | Mandarin | PosSim99.02 | 8 | |
| Cross-gender Retrieval | Mandarin M→F | Top-199.27 | 8 | |
| Automatic Speech Recognition | Mandarin | CER0.73 | 7 | |
| Singing Voice Synthesis | Mandarin (test) | LSD2.364 | 5 | |
| Language Modeling | Mandarin Tail (test) | Relative P95 RTF Reduction-36.23 | 3 | |
| Language Modeling | Mandarin STT (test) | Relative P95 RTF Reduction-30.21 | 3 | |
| Language Modeling | Mandarin VA (test) | Relative P95 RTF Reduction-2.11 | 3 | |
| Accent Shifting | Mandarin (evaluation) | Target Accent Probability33.8 | 2 | |
| Speech-to-Singing conversion | Mandarin (test) | LSD5.066 | 1 |
Showing 9 of 9 rows