Target Speaker Extraction on AISHELL Noisy zero-shot
10.2SI-SDRWHYV
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| WHYVTraining Dataset=Libri2Mix2024.10 | 10.2 | 10.1 | 2.6 | |
| WHYVTraining Dataset=WHAMR!2024.10 | 7.4 | 7.1 | 2.1 | |
| SpeakerBeamTraining Dataset=Libri2Mix2024.10 | 6.5 | 5.5 | 1.8 | |
| X-TF-GridnetTraining Dataset=WHAMR!2024.10 | 4.1 | 4 | 1.8 | |
| BSRNNTraining Dataset=VoxCeleb12024.10 | 2.8 | -1.4 | 0.6 |