I have been hunting for beamforming for a while and knew all was included in speechbrain but from previous tries with torchaudio failed with forced vendor libs of intel_mkl
A very nice chap has provide all and instructions for RaspiOSLite64
https://mathinf.eu/pytorch/arm64/2021-01/
So yeah for the first time just managed to install Speechbrain on a Pi3
I will let you know how I go on with the the beamformer and publish load.
It installs the arm-compute lib and a whole rake of others but just follow the install then pip install speechbrain and I can confirm the install works but here goes trying some beamformers
[EDIT]
(venv) pi@raspberrypi:~/speech-brain/speechbrain-0.5.7 $ python3 delay-sum.py
/home/pi/speech-brain/venv/lib/python3.7/site-packages/torch/functional.py:585: UserWarning: stft with return_complex=False is deprecated. In a future pytorch release, stft will return complex tensors for all inputs, and return_complex=False will raise an error.
Note: you can still call torch.view_as_real on the complex output to recover the old return format. (Triggered internally at ../aten/src/ATen/native/SpectralOps.cpp:483.)
normalized, onesided, return_complex)
Traceback (most recent call last):
File "delay-sum.py", line 38, in <module>
Xs = stft(xs)
File "/home/pi/speech-brain/venv/lib/python3.7/site-packages/torch/nn/modules/module.py", line 880, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/pi/speech-brain/speechbrain-0.5.7/speechbrain/processing/features.py", line 171, in forward
return_complex=False,
File "/home/pi/speech-brain/venv/lib/python3.7/site-packages/torch/functional.py", line 585, in stft
normalized, onesided, return_complex)
RuntimeError: fft: ATen not compiled with MKL support
Its not speechbrain and unsure why pytorch persists to force vendor specific libs for certain functions and still haven’t found a complete 100% Arm64 install.