Audio-Visual Speech Recognition on LRS2 Music noise (CER Multi-SNR)
16.64CER (-10 dB)VisG AV-HuBERT
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VisG AV-HuBERTArchitecture=Large2026.04 | 16.64 | 12.38 | 10.64 | 10.06 | 9.86 | |
| AV-HuBERTArchitecture=Large2026.04 | 17.4 | 12.9 | 11.11 | 10.38 | 10.23 | |
| VisG AV-HuBERTArchitecture=Base2026.04 | 20.43 | 14.97 | 11.64 | 10.31 | 9.76 | |
| AV-HuBERTArchitecture=Base2026.04 | 21.25 | 14.36 | 11.55 | 9.94 | 9.24 |