Child Automatic Speech Recognition on MyST 5h
12.88WERWavLM + ΔW2V2
Evaluation Results
| Method | Links | |
|---|---|---|
| WavLM + ΔW2V2Fusion Method=Concat2026.01 | 12.88 | |
| WavLM + ΔHuBERTFusion Method=Concat2026.01 | 12.96 | |
| WavLMModel Type=Single fine-tuned SSL2026.01 | 13.27 | |
| HuBERTModel Type=Single fine-tuned SSL2026.01 | 14.67 | |
| W2V2Model Type=Single fine-tuned SSL2026.01 | 15.65 |