Speaker-attributed Automatic Speech Recognition on AISHELL-4
21.36WERGLSC-SDR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GLSC-SDR2026.03 | 21.36 | 23.49 | 2.13 | 90.96 | |
| ViebVoice-ASR2026.03 | 21.4 | 24.99 | 3.59 | — | |
| Gemini-2.5-pro*source=converted from [18]2026.03 | 22.42 | 31.59 | 9.17 | — | |
| Gemini-3-pro*source=converted from [18]2026.03 | 22.75 | 27.43 | 4.68 | — | |
| Qwen2.5-omni-sftmode=fine-tuned2026.03 | 23.83 | 26.34 | 2.51 | 89.93 | |
| Qwen2.5-omni2026.03 | 31.46 | 46.28 | 14.82 | 76.68 |