Speaker-Attributed Automatic Speech Recognition on Candor Global Meeting-level
17.67DERPipeline
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PipelineInference Setting=Non-streaming2026.03 | 17.67 | 17.62 | |
| SortformerInference Setting=Non-streaming2026.03 | 18.03 | — | |
| G-STARInference Setting=Non-streaming2026.03 | 24.89 | 15.17 | |
| Vibevoice-ASRInference Setting=Non-streaming2026.03 | 25.68 | 27.24 | |
| ParakeetInference Setting=Non-streaming2026.03 | — | 26.92 |