Multi-speaker speech recognition on WSJ0-2mix 16 kHz (test)
14.9WERConditional-Chain-Transformer
Evaluation Results
| Method | Links | |
|---|---|---|
| Conditional-Chain-TransformerSystem=CTC, Training Data=WSJ0-1&2&3mix2020.06 | 14.9 | |
| DPCL + DNN-HMMSystem=HMM, Training Data=WSJ0-2mix2020.06 | 16.5 | |
| PIT-TransformerSystem=Attention-based, Training Data=WSJ0-2mix2020.06 | 17.7 | |
| Conditional-Chain-TransformerSystem=CTC, Training Data=WSJ0-2mix2020.06 | 24.7 | |
| PIT-RNNSystem=Attention-based, Training Data=WSJ0-2mix2020.06 | 25.4 | |
| PIT-DNN-HMMSystem=HMM, Training Data=WSJ0-2mix2020.06 | 28.2 | |
| DPCL + GMM-HMMSystem=HMM, Training Data=WSJ0-2mix2020.06 | 30.8 | |
| PIT-TransformerSystem=CTC, Training Data=WSJ0-2mix2020.06 | 31.2 |