Audio Visual Speech Recognition On Lrs3 Ted
评估指标
Word Error Rate (WER)
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | Word Error Rate (WER) |
|---|---|
| discriminative-multi-modality-speech | 6.8 |
| audio-visual-representation-learning-via | 1.3 |
| deep-audio-visual-speech-recognition | 7.2 |
| recurrent-neural-network-transducer-for-audio | 4.5 |
| end-to-end-audio-visual-speech-recognition | 2.3 |
| robust-self-supervised-audio-visual-speech | 1.4 |
| large-language-models-are-strong-audio-visual | 0.77 |
| auto-avsr-audio-visual-speech-recognition | 0.9 |
| whisper-flamingo-integrating-visual-features | 0.76 |
| jointly-learning-visual-and-auditory-speech | 1.4 |
| zero-avsr-zero-shot-audio-visual-speech | 1.5 |
| mms-llama-efficient-llm-based-audio-visual-1 | 0.74 |