Average Jaccard Similarity on Noisy Speech Utterances
0.72Avg Jaccard Sim (-5dB)BEST-STD 2.0
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BEST-STD 2.0Type=Speech Tokens, Tokens=10242025.12 | 0.72 | 0.78 | 0.81 | 0.83 | 0.84 | |
| Ours - TransformerType=Speech Tokens, Tokens=10242025.12 | 0.67 | 0.73 | 0.75 | 0.77 | 0.77 | |
| WavLM-LargeType=ASR Posteriors, Tokens=322025.12 | 0.62 | 0.67 | 0.7 | 0.71 | 0.71 | |
| HuBERT-LargeType=ASR Posteriors, Tokens=322025.12 | 0.46 | 0.59 | 0.67 | 0.71 | 0.72 | |
| BEST-STDType=Speech Tokens, Tokens=10242025.12 | 0.21 | 0.29 | 0.42 | 0.6 | 0.65 | |
| WavLM-LargeType=Speech Tokens, Tokens=10002025.12 | 0.18 | 0.19 | 0.21 | 0.22 | 0.23 | |
| SpeechTokenizerType=Speech Tokens, Tokens=10242025.12 | 0.09 | 0.12 | 0.15 | 0.18 | 0.19 |