Speaker-Attributed Automatic Speech Recognition on CallHome (test)
2.1WDERSAA+SID
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SAA+SIDTraining Mode=Best SAA+SID configuration2026.04 | 2.1 | 17.8 | |
| NemoDiarization Model=titanet_large, ASR Model=stt_en_conformer_ctc_large2026.04 | 7.1 | 20.3 | |
| PyAnnote+GraniteDiarization Tool=pyannote.audio-3.1, ASR Tool=Granite speech2026.04 | 15.1 | 42.5 | |
| PyAnnote+WhisperDiarization Tool=pyannote.audio-3.1, ASR Tool=Whisper (large v3)2026.04 | 17.1 | 66.6 |