Child Automatic Speech Recognition on MyST Full
9.64WERWavLM + ΔW2V2
Evaluation Results
| Method | Links | |
|---|---|---|
| WavLM + ΔW2V2Fusion Method=Concat2026.01 | 9.64 | |
| WavLM + ΔHuBERTFusion Method=Concat2026.01 | 9.71 | |
| WavLMModel Type=Single fine-tuned SSL2026.01 | 10.16 | |
| W2V2Model Type=Single fine-tuned SSL2026.01 | 10.96 | |
| HuBERTModel Type=Single fine-tuned SSL2026.01 | 11.04 |