Speech Recognition on GigaSpeech DEV
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 | SAMBA ASR | 9.12 | No | Samba-ASR: State-Of-The-Art Speech Recognition L... | 2025-01-06 | - |
| 2 | Zipformer+pruned transducer w/ CR-CTC (no external language model) | 9.95 | No | CR-CTC: Consistency regularization on CTC for im... | 2024-10-07 | Code |
| 3 | Zipformer+pruned transducer (no external language model) | 10.09 | No | CR-CTC: Consistency regularization on CTC for im... | 2024-10-07 | Code |
| 4 | Zipformer+CR-CTC (no external language model) | 10.15 | No | CR-CTC: Consistency regularization on CTC for im... | 2024-10-07 | Code |
| 5 | Conformer/Transformer-AED | 10.9 | No | GigaSpeech: An Evolving, Multi-domain ASR Corpus... | 2021-06-13 | Code |