How to make the Separator use the CPU while the GPU can be used

[RDGR8]

I want give the user the choice to use the CPU or the GPU, how could I do that? Is there an argument for that?

[DJVova]

I have a similar question. I have several GPUs in my system. How can I select the necessary GPU for processing?

[Lixi20]

if you want use GPU: pip install "audio-separator[gpu]"
if you want use CPU: pip install "audio-separator[cpu]"

audio-separator[gpu] uses GPU by default

[beveradb]

There isn't currently a parameter to control this, it always tries to use GPU acceleration if available as it's assumed that will be faster.
I don't really understand the use case for not using the GPU if there's one available, but feel free to raise a PR adding a parameter to disable GPU if you want.