Speech Separation On Wsj0 5Mix
Métriques
SI-SDRi
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | SI-SDRi |
|---|---|
| toward-speech-separation-in-the-pre-cocktail | 11.14 |
| voice-separation-with-an-unknown-number-of | 10.56 |
| boosting-unknown-number-speaker-separation | 21.0 |
| many-speakers-single-channel-speech | 13.22 |
| sequence-to-multi-sequence-learning-via | 11.7 |
| multi-decoder-dprnn-high-accuracy-source | 5.9 |