Speech Enhancement on WSJ0 + WHAM! SNR [-8, 0] dB
1.27PESQRemixIT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| RemixITModel=DPRNN*, Sequential Scene?=Sequential, Adaptable Params (#)=89,258, Adaptable Params (%)=1002026.03 | 1.27 | 74.51 | 5.82 | |
| Self-supervised Low-Rank AdaptationModel=DPRNN*, Sequential Scene?=Sequential, Adaptable Params (#)=708, Adaptable Params (%)=0.792026.03 | 1.27 | 75.18 | 5.85 | |
| Self-supervised Low-Rank AdaptationModel=DPRNN*, Sequential Scene?=Isolated, Adaptable Params (#)=708, Adaptable Params (%)=0.792026.03 | 1.26 | 75.52 | 5.52 | |
| RemixITModel=DPRNN*, Sequential Scene?=Isolated, Adaptable Params (#)=89,258, Adaptable Params (%)=1002026.03 | 1.24 | 73.74 | 5.58 | |
| Self-supervised Low-Rank AdaptationModel=GRU*, Sequential Scene?=Sequential, Adaptable Params (#)=512, Adaptable Params (%)=0.222026.03 | 1.23 | 72.65 | 4.84 | |
| Self-supervised Low-Rank AdaptationModel=GRU*, Sequential Scene?=Isolated, Adaptable Params (#)=512, Adaptable Params (%)=0.222026.03 | 1.22 | 71.96 | 4.71 | |
| PretrainedModel=DPRNN*2026.03 | 1.21 | 74.93 | 5.15 | |
| RemixITModel=GRU*, Sequential Scene?=Isolated, Adaptable Params (#)=230,144, Adaptable Params (%)=1002026.03 | 1.19 | 71.64 | 4.64 | |
| RemixITModel=GRU*, Sequential Scene?=Sequential, Adaptable Params (#)=230,144, Adaptable Params (%)=1002026.03 | 1.18 | 70.34 | 4.63 | |
| PretrainedModel=GRU*2026.03 | 1.16 | 71.01 | 3.86 | |
| Noisy Speech2026.03 | 1.08 | 67.88 | -3.89 |