Full-Attention Residual Network for End-to-End Speech Separation The code will be shared if the submission is accepted! Platform CUDA 11.1 cuDNN 8.0 PyTorch 1.7