Speech Separation on WSJ0-5mix (test)
11.7SI-SNRiConditional TasNet*
Evaluation results
| Method | Links | |
|---|---|---|
| Conditional TasNet*Training Data=WSJ0 2-5 mix, Training with variable number of speakers=true, Architecture=single2020.06 | 11.7 | |
| Voice SeparationTraining Data=WSJ0-Nmix, Training with variable number of speakers=false, Architecture=multiple2020.06 | 10.6 | |
| DPRNNTraining Data=WSJ0-Nmix, Training with variable number of speakers=false, Architecture=multiple2020.06 | 8.4 | |
| ConvTasNetTraining Data=WSJ0-Nmix, Training with variable number of speakers=false, Architecture=multiple2020.06 | 6.8 |