Speech Separation On Wsj0 4Mix
評価指標
SI-SDRi
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | SI-SDRi | Paper Title | Repository |
|---|---|---|---|
| Conditional TasNet | 12.5 | Sequence to Multi-Sequence Learning via Conditional Chain Mapping for Mixture Signals | - |
| SepTDA | 22.0 | Boosting Unknown-number Speaker Separation with Transformer Decoder-based Attractor | - |
| OR-PIT | 10.2 | - | - |
| Multi-Decoder DPRNN | 9.3 | Multi-Decoder DPRNN: High Accuracy Source Counting and Separation | |
| Gated DualPathRNN | 12.88 | Voice Separation with an Unknown Number of Multiple Speakers |
0 of 5 row(s) selected.