machine learning on mac mini m4

[pippo73]

Hello all,
I would like to know if there is any guide in order to use all the power of a mac mini m4. How should I configure it in order to use all the power of this small beast?
tnx for any hint pointing me to the right direction.

[bo0tzz]

I don't think there is a way to do this, my understanding is that ARM mac hardware acceleration can't be passed through to a container.

[pippo73]

I don't have the mac on my hands now, but judging by some articles I've found in the net seems that some kind of collaboration between docker and gpu is possibile
https://medium.com/@andreask_75652/gpu-accelerated-containers-for-m1-m2-m3-macs-237556e5fe0b
this is one of the articles I've found. Maybe it can help you too ;-)

[mmomjian]

Hello, we strongly recommend the use of Linux for Immich. Docker does not work very well on Windows and Mac.

[pippo73]

ok I will do some tests as soon I willl get the machine and then I will be back

[mertalev]

I don't have the mac on my hands now, but judging by some articles I've found in the net seems that some kind of collaboration between docker and gpu is possibile

https://medium.com/@andreask_75652/gpu-accelerated-containers-for-m1-m2-m3-macs-237556e5fe0b

this is one of the articles I've found. Maybe it can help you too ;-)

This is an interesting read, but it's pretty hacky and much worse than native performance. If we support CoreML, it'll more than likely be through running the ML service natively outside of Docker.

[CharlesCCC]

Similar from #10454

I'd believe we can achieve it within docker, we might need run the machine learninig service manually outside docker.

[pippo73]

I've joined the tread you indicated me.
Actually I've installed podman on my macminim4 and everything is up and running. But I will be glad to have a native solution to squeeze all the power of it.
How can I see how it works? are any commands I can run to test it?