Automatic Speech Recognition on SUPERB 48kHz
5.53WERRe. 16kHz HuBERT Base
Evaluation Results
| Method | Links | |
|---|---|---|
| Re. 16kHz HuBERT BaseFine-tuning protocol=Mixed-rate2026.03 | 5.53 | |
| MSRHuBERTFine-tuning protocol=Mixed-rate2026.03 | 5.56 | |
| Re. 16kHz HuBERT BaseFine-tuning protocol=Single-rate2026.03 | 5.61 | |
| Re. 24kHz HuBERT BaseFine-tuning protocol=Single-rate2026.03 | 5.7 | |
| Re. 24kHz HuBERT BaseFine-tuning protocol=Mixed-rate2026.03 | 5.74 | |
| MSRHuBERTFine-tuning protocol=Single-rate2026.03 | 5.83 | |
| Re. 48kHz HuBERT BaseFine-tuning protocol=Single-rate2026.03 | 5.95 | |
| HuBERT BaseFine-tuning protocol=Single-rate2026.03 | 5.96 | |
| Re. 48kHz HuBERT BaseFine-tuning protocol=Mixed-rate2026.03 | 6.11 | |
| HuBERT BaseFine-tuning protocol=Mixed-rate2026.03 | 6.63 | |
| Re. 16kHz HuBERT Base (w/o resampling)Fine-tuning protocol=Mixed-rate, Resampling in fine-tuning=No2026.03 | 33.72 | |
| Re. 16kHz HuBERT Base (w/o resampling)Fine-tuning protocol=Single-rate, Resampling in fine-tuning=No2026.03 | 38.18 |