Closed Tmatz27 closed 6 months ago
Can you post the docker logs ? You should be able to access it with docker logs Immich
I have included log files before i reran facial detection. After I ran facial detection again i recieved so many errors I could not see the whole log file in the console. The main statement is captured in a screen shot attached to this comment as well. The error just keeps repeating with different "id"s. There is no config file in the folder so not sure how correct these errors. Screenshot 2024-03-22 105809 root@Tower~# docker logs immich.txt
Why is your machine url set to 127.0.0.0:443
?
Have you changed something in the admin settings / template ?
Nope, the rest of the docker is working correctly. Ill attach my setup. If there was a config file i think i would be able to fix this.
What's the machine learning URL in administration > settings > Machine Learning settings ?
Wow i cant believe i missed that! it was set to 127.0.0.0. Im assuming i need to set this to my unraid server of 192.168.1.132 correct? It only has an intel igpu though. I have two other PCs with nvidia gpus so i could set it to those ip addresses?
No just hit reset to default
and save
.
Immich does not support multiple machine learning instances OOTB, but you can have it working by setting multiple immich_machine_learning containers and a load balancer. Really not straightforward but achievable.
Not sure how that setting got changed but can confirm that facial recognition is currently working. Thank you for the help!
I do have one follow up question. Im looking at the ML docs on the Immich page. It references inline adding the hardware acceleration config into the docker-compose.yml file but this is not present in the Unraid immich file system. Ive gone through every single folder and i can not find it
Yeah that's how you're supposed to deploy Immich with the "official" way (with docker-compose).
All information related to the deployment of this AIO image, including machine learning acceleration, can be found in the README of this repo.
I'm not able to get machine learning or facial recognition working in unraid. I've got a tons of pictures with people in them and no matter how many times I run the facial recognition it always comes back empty. It completes the job in seconds even though it's going through thousands of photos. Not sure what to do, I can't find a log file or config file in the immich app data folder either so I can't post logs