Phoneme Recognition on LibriSpeech clean 100h noise-augmented (test)
5.11PERSSL-SoftDTW
Evaluation results
| Method | Links | |
|---|---|---|
| SSL-SoftDTWSSL Fine-tuning of SE Model=SSL-SoftDTW, Enhancement=true2026.01 | 5.11 | |
| SSL-MSESSL Fine-tuning of SE Model=SSL-MSE, Enhancement=true2026.01 | 5.17 | |
| SSL-MSE-PADSSL Fine-tuning of SE Model=SSL-MSE-PAD, Enhancement=true2026.01 | 5.18 | |
| master64SSL Fine-tuning of SE Model=None, Enhancement=true2026.01 | 5.28 | |
| HuBERT (No Enhancement)SSL Fine-tuning of SE Model=None, Enhancement=false2026.01 | 5.55 |