| Benchmark | Model | Rank | Results |
|---|---|---|---|
| audio-visual-speech-recognition-on-lrs2 | End2end Conformer | #4 | Test WER: 3.7 |
| audio-visual-speech-recognition-on-lrs3-ted | Hyb-Conformer | #9 | Word Error Rate (WER): 2.3 |
| automatic-speech-recognition-on-lrs2 | End2end Conformer | #4 | Test WER: 3.9 |
| lipreading-on-lrs2 | Hybrid CTC / Attention | #8 | Word Error Rate (WER): 39.1 |
| lipreading-on-lrs3-ted | Hyb + Conformer | #13 | Word Error Rate (WER): 43.3 |