Speech Quality Assessment (MOS Prediction) on Tencent w/o R (test)
0.751MSEbaseline
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| baselinetrain data=NISQA+AudioMOS_train2026.02 | 0.751 | 0.917 | 0.901 | |
| baselinetrain data=AudioMOS_train2026.02 | 1.002 | 0.691 | 0.687 | |
| SA-SSL-MOStrain data=AudioMOS_train2026.02 | 1.097 | 0.669 | 0.666 | |
| SA-SSL-MOStrain data=NISQA+AudioMOS_train2026.02 | 1.192 | 0.877 | 0.891 |