Speech Quality Assessment (MOS Prediction) on Tencent w R (test)
0.421MSEbaseline
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| baselinetrain data=NISQA+AudioMOS_train2026.02 | 0.421 | 0.814 | 0.78 | |
| SA-SSL-MOStrain data=NISQA+AudioMOS_train2026.02 | 0.453 | 0.795 | 0.741 | |
| baselinetrain data=AudioMOS_train2026.02 | 0.577 | 0.638 | 0.542 | |
| SA-SSL-MOStrain data=AudioMOS_train2026.02 | 0.712 | 0.63 | 0.544 |