Multi-speaker ASR on LibriSpeech3Mix
6.2WERSOT-Conformer [46]
Evaluation results
| Method | Links | |
|---|---|---|
| SOT-Conformer [46]Training Set=LibriSpeechMix, Training hours=960 hours, Mix Type=speech only, Overlap Mode=partial2025.08 | 6.2 | |
| Whisper-medium-SS-TTI [47]Training Set=LibriSpeechMix, Training hours=960 hours, Mix Type=speech only, Overlap Mode=partial2025.08 | 7.5 | |
| UME ASR init. (0.33, 0.33, 0.34) w/ RWSE (A4)Training Set=LibriMix, Training hours=460 hours, Mix Type=mixclean: speech only, Overlap Mode=max, Initialization=ASR init., Training weights (λasr, λdiar, λsep)=(0.33, 0.33, 0.34), RWSE=true2025.08 | 12.9 | |
| UME (0.33, 0.33, 0.34) + ASR init.Training Set=LibriMix, Training hours=460 hours, Mix Type=mixboth: speech + noise, Overlap Mode=max, Initialization=ASR init., Training weights (λasr, λdiar, λsep)=(0.33, 0.33, 0.34)2025.08 | 15.7 | |
| UME (λasr = 1.0)Training Set=LibriMix, Training hours=460 hours, Mix Type=mixboth: speech + noise, Overlap Mode=max, Training weights (λasr, λdiar, λsep)=(1.0, 0, 0)2025.08 | 16 | |
| UME ASR init. (0.33, 0.33, 0.34) w/ RWSE (A3)Training Set=LibriMix, Training hours=460 hours, Mix Type=mixboth: speech + noise, Overlap Mode=max, Initialization=ASR init., Training weights (λasr, λdiar, λsep)=(0.33, 0.33, 0.34), RWSE=true2025.08 | 18 | |
| UME (0.1, 0.1, 0.8) + ASR init.Training Set=LibriMix, Training hours=460 hours, Mix Type=mixboth: speech + noise, Overlap Mode=max, Initialization=ASR init., Training weights (λasr, λdiar, λsep)=(0.1, 0.1, 0.8)2025.08 | 20.3 | |
| SOT [10]Training Set=LibriSpeechMix, Training hours=960 hours, Mix Type=speech only, Overlap Mode=partial2025.08 | 24 | |
| PIT LSTM-AED [10]Training Set=LibriSpeechMix, Training hours=960 hours, Mix Type=speech only, Overlap Mode=partial2025.08 | 52.3 |