Feature Request: Add Classic Speech Separation Methods such as Conv-TasNet and DualPathRNN

[JusperLee]

Conv-TasNet and DualPathRNN have proven effective in handling speech separation tasks and can be used for various applications. Few repositories are currently available for separation solutions in hardware devices.

Could this feature be added soon, or do you need any help implementing it?

[StuartIanNaylor]

I have seen your repo's @JusperLee and second they also world be a great option and hope you do.

I still think multi-channel help more with reverberation and often post better results for lower load. https://github.com/fakufaku/torchiva Has some great examples where maybe you place KWS/VAD on the output to select the voice channel even though maybe luck it seems channel 0 always (Thought it would be random)