ASR is all you need: cross-modal distillation for lip reading
Open paper
Benchmark
Model
Rank
Results
lipreading-on-lrs2
CTC + KD ASR
–
Word Error Rate (WER): 53.2
lipreading-on-lrs3-ted
CTC + KD
–
Word Error Rate (WER): 59.8
Rank counts only results with a code link.