My poor Opteron with 8 GB was straining a lot with the transcoding despite the hardware acceleration…
I guess that tells me how well the ML facial recognition would work
I’ve disabled it for now; need to tweak some other things first. And maybe I can find a used piece of DDR4 ECC RAM.
@ah I haven’t tried it yet but maybe I’ll run an experiment soon with a small library…
(But also, this: https://mastodon.hupel.info/@lars/116234703149041761)
@[email protected] I’m currently running it with docker compose. Next step would be to switch to rootless (the root containers irk me). Then setting up proper backup, then tweaking Redis, then importing all my old images, … 😂