This is the official GitHub repo for ""CasNet: Investigating Channel Robustness for Speech Separation"".
The code is mainly built on the SpeechBrain toolkit. For the detailed instructions, please go to the official page of SpeechBrain. For now, the complete code is in our forked speechbrain https://github.com/Sinica-SLAM/speechbrain/tree/develop/recipes/channel_adapt/separation. We are still working on moving the structure to an independent place.