Speaker-aware ASR on AISHELL-1 & AISHELL-2 augmented with VoxCeleb & MUSAN (2 dB SNR)
5.34WERUAF-30B-A3B
Evaluation results
| Method | Links | |
|---|---|---|
| UAF-30B-A3BReference prompt=true, Enrollment duration=5-second2026.04 | 5.34 | |
| Kimi-Audio-7BReference prompt=true, Enrollment duration=5-second2026.04 | 36.25 | |
| Qwen3-Omni-30B-A3BReference prompt=true, Enrollment duration=5-second2026.04 | 38.6 | |
| Qwen2.5-Omni-7BReference prompt=true, Enrollment duration=5-second2026.04 | 81.77 |