Er loopt een gast rond die CD's van "zijn" metal band verkoopt. Aardige gast. Maar verder is het een scammer. Google Defrage Reload als je meer wilt weten.
I'm having problems getting machine learning to work properly.
My setup currently:
Raspberry pi 3 running Immich without the remote-machine-learning container.
Game pc running only the remote-machine-learning container which I aim start manually every now and then to queue missing images.
This seems to work fine. Immich is able to offload ML tasks to the PC. I see GPU usage spikes and see lots of logs about new faces etc in the Immich container. With the PC turned off I can search on these faces etc. Duplicate detection is working as well.
However, searching on context does not work at all. It logs errors about the ML url not being reachable - but why would it need to access the ML container to search through cached results?
Am I misunderstanding how the context search is supposed to work?
I did try resetting the ML settings (and then restore the ip to my PC) and requeue all images, but that didn't work either.