MPS support for Torchaudio?

Hi everyone,

My question is simple: Are Torchaudio APIs soon gonna be able to support MPS to perform computations on M1 mac GPUs?

Here is the doc and it doesn’t mention any support thus far, apart from CPU and CUDA:
https://pytorch.org/audio/main/supported_features.html

Is there anything in the pipes with a coming release?

Cheers

Antoine