Automatic Speech Recognition on Common Voice Mandarin (test)
17.82CERWSM
Evaluation results
| Method | Links | |
|---|---|---|
| WSMSSL Model=MMS, Version=Ours2026.02 | 17.82 | |
| MMSSSL Model=MMS, Version=Baseline2026.02 | 18.28 | |
| WSMSSL Model=mHuBERT, Version=Ours2026.02 | 18.99 | |
| mHuBERTSSL Model=mHuBERT, Version=Baseline2026.02 | 19.82 | |
| WSMSSL Model=XLS-R, Version=Ours2026.02 | 19.98 | |
| XLS-RSSL Model=XLS-R, Version=Baseline2026.02 | 20.01 | |
| WSMSSL Model=HuBERT, Version=Ours2026.02 | 21.56 | |
| WSMSSL Model=data2vec, Version=Ours2026.02 | 21.96 | |
| HuBERTSSL Model=HuBERT, Version=Baseline2026.02 | 22.02 | |
| WSMSSL Model=wav2vec 2.0, Version=Ours2026.02 | 22.85 | |
| wav2vec 2.0SSL Model=wav2vec 2.0, Version=Baseline2026.02 | 24.1 | |
| data2vecSSL Model=data2vec, Version=Baseline2026.02 | 24.43 |