Speech Recognition on WSJ eval93
Metric: Word Error Rate (WER) (lower is better)
Metric: Word Error Rate (WER) (lower is better)
| # | Model↕ | Word Error Rate (WER)▲ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | Deep Speech 2 | 4.98 | Yes | Deep Speech 2: End-to-End Speech Recognition in ... | 2015-12-08 | Code |
| 2 | CTC-CRF 4gram-LM | 6.23 | No | - | - | Code |
| 3 | Convolutional Speech Recognition | 6.8 | No | Fully Convolutional Speech Recognition | 2018-12-17 | - |