Can go to sleep now . Got #immich and #headscale / #tailscale working. I am using a #hetzner arm server . For 5 Euros a month ! Doing my first sync. Performance is outstanding ! Sadly I never thought that I require a gpu in my #homelab . Now I consider to buy a cpu with an igpu to make the machine learning part work. Can I run it without a gpu or does it make no sense ? #
@__sharky__ For what part of #immich a GPU should be needed? Running it on a 0815-VM, everything seems to be working including face recognition?
@wusel ok 🫡 , I thought it is a requirement to have a gpu for face recognition. The server is under full load now , I will check tomorrow if it works . Thank you
@__sharky__ Works for me like a charm, it's actually amazing how it even recognizes our dog over the years, as well as the kids; there seems to be a break around 6 years of age, but it you align them once, it's really awesome ;-)
@wusel Ok , machine learning is working fine after starting the container :) . I am really happy with the installation.
@wusel Sounds exactly what I was hoping for 👌 . Do you run it via docker compose ? I am asking because I am not running the machine learning container at the moment and the load dropped to 2% after one or two hours. So it seems there isn’t anything going on anymore. Pretty sure I have to enable it to get those features , but something for tomorrow 🛌 😴