From: Time-domain adaptive attention network for single-channel speech separation
Model
SI-SNRi
SDRi
DPRNN
14.7
15.5
DPTNet
16.3
TAANet
16.0
16.8