| TEVR: Improving Speech Recognition by Token Entropy Variance Reduction | ✓ Link | 3.64% | 1.54% | wav2vec 2.0 XLS-R 1B + TEVR (5-gram) | 2022-06-25 |
| TEVR: Improving Speech Recognition by Token Entropy Variance Reduction | ✓ Link | 3.70% | | wav2vec 2.0 XLS-R 1B + TEVR (4-gram) | 2022-06-25 |
| Scribosermo: Fast Speech-to-Text models for German and other Languages | ✓ Link | 4.05% | 1.37% | ConformerCTC-L (5-gram) | 2021-10-15 |
| NeMo: a toolkit for building AI applications using Neural Modules | ✓ Link | 4.09% | | canary-1b-flash | 2025-03-07 |
| TEVR: Improving Speech Recognition by Token Entropy Variance Reduction | ✓ Link | 4.38% | 1.62% | wav2vec 2.0 XLS-R 1B (5-gram) | 2022-06-25 |
| NeMo: a toolkit for building AI applications using Neural Modules | ✓ Link | 6.03% | | ConformerCTC-L (4-gram) | 2019-09-14 |
| Automatic Speech Recognition in German: A Detailed Error Analysis | | 6.28% | | Conformer Transducer (no LM) | 2022-08-03 |
| Robust Speech Recognition via Large-Scale Weak Supervision | ✓ Link | 6.4% | | Whisper (Large v2) | 2022-12-06 |
| Scribosermo: Fast Speech-to-Text models for German and other Languages | ✓ Link | 6.6% | 2.7% | QuartzNet15x5DE (D37, 5-gram) | 2021-10-15 |
| NeMo: a toolkit for building AI applications using Neural Modules | ✓ Link | 6.68% | | ConformerCTC-L (no LM) | 2019-09-14 |
| Scribosermo: Fast Speech-to-Text models for German and other Languages | ✓ Link | 7.33% | 2.05% | ConformerCTC-L (no LM) | 2021-10-15 |
| Scribosermo: Fast Speech-to-Text models for German and other Languages | ✓ Link | 7.7% | 3.2% | QuartzNet15x5DE (CV-only, 5-gram) | 2021-10-15 |
| VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretation | ✓ Link | 7.8% | | VoxPopuli (n-gram) | 2021-01-02 |
| TEVR: Improving Speech Recognition by Token Entropy Variance Reduction | ✓ Link | 10.10% | | wav2vec 2.0 XLS-R 1B + TEVR (no LM) | 2022-06-25 |
| TEVR: Improving Speech Recognition by Token Entropy Variance Reduction | ✓ Link | 12.06% | | wav2vec 2.0 XLS-R (no LM) | 2022-06-25 |