| Paper | Code | Character Error Rate (CER) | ModelName | ReleaseDate |
|---|---|---|---|---|
| FunASR: A Fundamental End-to-End Speech Recognition Toolkit | ✓ Link | 6.97 | Paraformer-large | 2023-05-18 |
| 3M: Multi-loss, Multi-path and Multi-level Neural Networks for speech recognition | ✓ Link | 7.19 | Conformer-MoE (64e) | 2022-04-07 |
| Zipformer: A faster and better encoder for automatic speech recognition | ✓ Link | 7.29 | Zipformer+pruned transducer (no external language model) | 2023-10-17 |
| 3M: Multi-loss, Multi-path and Multi-level Neural Networks for speech recognition | ✓ Link | 7.49 | Conformer-MoE (32e) | 2022-04-07 |
| 3M: Multi-loss, Multi-path and Multi-level Neural Networks for speech recognition | ✓ Link | 7.67 | Conformer-MoE (16e) | 2022-04-07 |
| WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognition | ✓ Link | 8.88 | Wenet | 2021-10-07 |
| WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognition | ✓ Link | 9.07 | Kaldi | 2021-10-07 |
| WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognition | ✓ Link | 9.7 | Espnet | 2021-10-07 |