Immich•10mo ago

Trying to get immich remote machine learning working.

I've set up immich in my synology nas and remote machine learning windows via docker desktop app. however the nas is not sending anything. really no idea how to get it fix. Anyone can help?

41 Replies

Immich•10mo ago

:wave: Hey @Dididididi, Thanks for reaching out to us. Please carefully read this message and follow the recommended actions. This will help us be more effective in our support effort and leave more time for building Immich :immich:. References - Container Logs: docker compose logs docs - Container Status: docker ps -a docs - Reverse Proxy: https://immich.app/docs/administration/reverse-proxy - Code Formatting https://support.discord.com/hc/en-us/articles/210298617-Markdown-Text-101-Chat-Formatting-Bold-Italic-Underline#h_01GY0DAKGXDEHE263BCAYEGFJA Checklist I have... 1. :ballot_box_with_check: verified I'm on the latest release(note that mobile app releases may take some time). 2. :ballot_box_with_check: read applicable release notes. 3. :ballot_box_with_check: reviewed the FAQs for known issues. 4. :ballot_box_with_check: reviewed Github for known issues. 5. :ballot_box_with_check: tried accessing Immich via local ip (without a custom reverse proxy). 6. :ballot_box_with_check: uploaded the relevant information (see below). 7. :ballot_box_with_check: tried an incognito window, disabled extensions, cleared mobile app cache, logged out and back in, different browsers, etc. as applicable (an item can be marked as "complete" by reacting with the appropriate number) Information In order to be able to effectively help you, we need you to provide clear information to show what the problem is. The exact details needed vary per case, but here is a list of things to consider: - Your docker-compose.yml and .env files. - Logs from all the containers and their status (see above). - All the troubleshooting steps you've tried so far. - Any recent changes you've made to Immich or your system. - Details about your system (both software/OS and hardware). - Details about your storage (filesystems, type of disks, output of commands like fdisk -l and df -h). - The version of the Immich server, mobile app, and other relevant pieces. - Any other information that you think might be relevant. Please paste files and logs with proper code formatting, and especially avoid blurry screenshots. Without the right information we can't work out what the problem is. Help us help you ;) If this ticket can be closed you can use the /close command, and re-open it later if needed.

DididididiOP•10mo ago

name: immich_remote_ml

services:
  immich-machine-learning:
    container_name: immich_machine_learning
    # For hardware acceleration, add one of -[armnn, cuda, openvino] to the image tag.
    # Example tag: ${IMMICH_VERSION:-release}-cuda
    image: ghcr.io/immich-app/immich-machine-learning:${IMMICH_VERSION:-release}-cuda
    extends:
      file: hwaccel.ml.yml
      service: cuda # set to one of [armnn, cuda, openvino, openvino-wsl] for accelerated inference - use the `-wsl` version for WSL2 where applicable
    volumes:
      - model-cache:/cache
    restart: always
    ports:
      - 3003:3003

volumes:
  model-cache:

name: immich_remote_ml

services:
  immich-machine-learning:
    container_name: immich_machine_learning
    # For hardware acceleration, add one of -[armnn, cuda, openvino] to the image tag.
    # Example tag: ${IMMICH_VERSION:-release}-cuda
    image: ghcr.io/immich-app/immich-machine-learning:${IMMICH_VERSION:-release}-cuda
    extends:
      file: hwaccel.ml.yml
      service: cuda # set to one of [armnn, cuda, openvino, openvino-wsl] for accelerated inference - use the `-wsl` version for WSL2 where applicable
    volumes:
      - model-cache:/cache
    restart: always
    ports:
      - 3003:3003

volumes:
  model-cache:

windows desktop docker app

DididididiOP•10mo ago

appreciate your help @sogan, let me know if u need anything more

mertalev•10mo ago

It looks like it’s being used based on the logs?

DididididiOP•10mo ago

dont think so

mertalev•10mo ago

It’s clearly getting requests since it’s loading models

DididididiOP•10mo ago

this was so long ago

DididididiOP•10mo ago

is it? im not sure or should i turn on more detail logs to see it better

mertalev•10mo ago

Unless you queried the remote ML server yourself, the only way you’d get those logs is if it got requests to process images

DididididiOP•10mo ago

I cant tell seems like server is not requesting to process anything

mertalev•10mo ago

You can check nvidia-smi or better yet nvtop to confirm GPU utilization

DididididiOP•10mo ago

my server is running stuff

mertalev•10mo ago

The remote ML server doesn’t output logs for individual requests fyi

DididididiOP•10mo ago

where do I input this nvtop

mertalev•10mo ago

For windows it might be easier to use nvidia-smi or a gui-based tool like hwinfo64 Also, if those ML logs are the last logs from the container, it means the remote container has been getting a steady stream of requests. It would have unloaded the models if they weren’t used within a 5 minute period

DididididiOP•10mo ago

ah i still cant confirm or do I need all normal container to get it working?

DididididiOP•10mo ago

I only have 1 atm

DididididiOP•10mo ago

perhaps it is working, i can see the gpu jumping around with network traffic

mertalev•10mo ago

Check the performance tab

arioch82•6mo ago

i am trying the same thing, i see similar things in the log, my GPU seems to be only used 7-8%, did you ever figure this out? looks like it's running but it could run a lot faster? my data is on a synology nas

arioch82•6mo ago

not sure if it's correct that it runs this every time: "loading detection model 'buffalo_l' to memory" i would expect the detection model to be loaded once and kept in memory while processing is happening?

arioch82•6mo ago

this is my config

name: immich_remote_ml

services:
  immich-machine-learning:
    container_name: immich_machine_learning
    # Note the `-cuda` at the end
    image: ghcr.io/immich-app/immich-machine-learning:${IMMICH_VERSION:-release}-cuda
    # Note the lack of an `extends` section
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities:
                - gpu
    volumes:
      - model-cache:/cache
    restart: always
    ports:
      - 3003:3003

volumes:
  model-cache:

name: immich_remote_ml

services:
  immich-machine-learning:
    container_name: immich_machine_learning
    # Note the `-cuda` at the end
    image: ghcr.io/immich-app/immich-machine-learning:${IMMICH_VERSION:-release}-cuda
    # Note the lack of an `extends` section
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities:
                - gpu
    volumes:
      - model-cache:/cache
    restart: always
    ports:
      - 3003:3003

volumes:
  model-cache:

those errors make me think there is something else going on there... [05/03/25 00:06:59] ERROR Worker (pid:4325) was sent code 139!

mertalev•6mo ago

It’s crashing with segmentation faults. That’s often a driver issue. Maybe check that docker desktop is up to date?

arioch82•6mo ago

i just installed it i did update the nvidia drivers right now after i started the container, let me kill it and restart it

mertalev•6mo ago

Yeah, updating the driver can also cause that

arioch82•6mo ago

same issue after killing the container and restarting it maybe i need to reboot

mertalev•6mo ago

You would probably need to restart docker Or restart your pc

arioch82•6mo ago

seems to be working without errors after the reboot, still only between 10-20% on the gpu tho

arioch82•6mo ago

even setting to 100 each doesn't seem to use the gpu much, spikes around 30%, are there any recommended values somewhere?

mertalev•6mo ago

Setting it that high is counterproductive. My guess is that it’s an IO bottleneck: you can’t load the images fast enough to keep the GPU saturated Are the thumbnails stored on an SSD or an HDD?

arioch82•6mo ago

hdd

mertalev•6mo ago

Yeah that would do it. Putting them on an SSD would make a huge diffference, not just in processing speed but also in day to day use

arioch82•6mo ago

i do want to add an ssd to the nas for caching... i am trying to find the docs for how to point immich thumbnails to that, do i just remap /thumbs in the volume?

mertalev•6mo ago

Yup https://immich.app/docs/guides/custom-locations

arioch82•6mo ago

thanks for all the help, i am gonna keep those values at 50 for now unless there is some recommended values per gpu somewhere 🙂

mertalev•6mo ago

I find that 16 is enough on a 4090 with an nvme. Also, try switching to a slower and better model. The default model is very fast but not as good as some others, so since you’re already bottlenecked by IO you may as well get more out of it

arioch82•6mo ago

which model do you use? i have a 3090

mertalev•6mo ago

ViT-SO400M-16-SigLIP2-384__webli right now, which is one of the best for english. There are charts on how the models compare here as well https://immich.app/docs/features/searching

Gaming

Programming

Trying to get immich remote machine learning working.

Did you find this page helpful?