| Benchmark | Model | Rank | Results |
|---|---|---|---|
| lipreading-on-lip-reading-in-the-wild | 3D Conv + ResNet-18 + MS-TCN + Multi-Head Visual-Audio Memory | #6 | Top-1 Accuracy: 88.5 |
| lipreading-on-lrs2 | Multi-head Visual-Audio Memory | #10 | Word Error Rate (WER): 44.5 |
| lipreading-on-lrw-1000 | 3D Conv + ResNet-18 + MS-TCN + Multi-Head Visual-Audio Memory | #3 | Top-1 Accuracy: 53.8 |