CM3588 armnn still showing CPUExecutionProvider

[binamov]

The bug

Hi folks,

Just set up immich on my CM3588. Followed the docs to configure armnn wherever needed, but logs are still showing CPUExecutionProvider.

Is there anything I can do to troubleshoot and fix?

The OS that Immich Server is running on

Debian bookworm

Version of Immich Server

v1.125.2

Version of Immich Mobile App

v.irrelevant

Platform with the issue

• Server
• Web
• Mobile

Your docker-compose.yml content

services:
  immich-server:
    container_name: immich_server
    image: ghcr.io/immich-app/immich-server:${IMMICH_VERSION:-release}
    extends:
      file: hwaccel.transcoding.yml
      service: rkmpp # set to one of [nvenc, quicksync, rkmpp, vaapi, vaapi-wsl] for accelerated transcoding
    volumes:
      # Do not edit the next line. If you want to change the media storage location on your system, edit the value of UPLOAD_LOCATION in the .env file
      - ${UPLOAD_LOCATION}:/usr/src/app/upload
      - /etc/localtime:/etc/localtime:ro
    env_file:
      - .env
    ports:
      - '2283:2283'
    depends_on:
      - redis
      - database
    restart: always
    healthcheck:
      disable: false
    labels:
      - homepage.group=media
      - homepage.name=immich
      - homepage.icon=immich.png
      - homepage.href=https://photos/
      - homepage.description=photos and videos

  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}-armnn
    extends: # uncomment this section for hardware acceleration - see https://immich.app/docs/features/ml-hardware-acceleration
      file: hwaccel.ml.yml
      service: armnn # set to one of [armnn, cuda, openvino, openvino-wsl] for accelerated inference - use the `-wsl` version for WSL2 where applicable
    volumes:
      - model-cache:/cache
    env_file:
      - .env
    restart: always
    healthcheck:
      disable: false

Your .env content

# machine learning
MACHINE_LEARNING_ANN=true

Reproduction steps

1. Configure armnn and rkmpp as described in docs
2. Look at logs of machine learning container

...

Relevant log output

[01/30/25 04:57:02] INFO     Starting gunicorn 23.0.0                           
[01/30/25 04:57:02] INFO     Listening at: http://[::]:3003 (9)                 
[01/30/25 04:57:02] INFO     Using worker: app.config.CustomUvicornWorker       
[01/30/25 04:57:02] INFO     Booting worker with pid: 10                        
[01/30/25 04:57:10] INFO     Started server process [10]                        
[01/30/25 04:57:10] INFO     Waiting for application startup.                   
[01/30/25 04:57:10] INFO     Created in-memory cache with unloading after 300s  
                             of inactivity.                                     
[01/30/25 04:57:10] INFO     Initialized request thread pool with 8 threads.    
[01/30/25 04:57:10] INFO     Application startup complete.                      
[01/30/25 05:57:06] INFO     Loading visual model 'ViT-B-32__openai' to memory  
[01/30/25 05:57:06] INFO     Setting execution providers to                     
                             ['CPUExecutionProvider'], in descending order of   
                             preference                                         
[01/30/25 05:57:07] INFO     Loading detection model 'buffalo_l' to memory      
[01/30/25 05:57:07] INFO     Setting execution providers to                     
                             ['CPUExecutionProvider'], in descending order of   
                             preference                                         
[01/30/25 05:57:08] INFO     Loading recognition model 'buffalo_l' to memory    
[01/30/25 05:57:08] INFO     Setting execution providers to                     
                             ['CPUExecutionProvider'], in descending order of   
                             preference                                         
[01/30/25 06:04:30] INFO     Shutting down due to inactivity.                   
[01/30/25 06:04:30] INFO     Shutting down                                      
[01/30/25 06:04:30] INFO     Waiting for application shutdown.                  
[01/30/25 06:04:31] INFO     Application shutdown complete.                     
[01/30/25 06:04:31] INFO     Finished server process [10]                       
[01/30/25 06:04:31] ERROR    Worker (pid:10) was sent SIGINT!                   
[01/30/25 06:04:31] INFO     Booting worker with pid: 966                       
[01/30/25 06:04:38] INFO     Started server process [966]                       
[01/30/25 06:04:38] INFO     Waiting for application startup.                   
[01/30/25 06:04:38] INFO     Created in-memory cache with unloading after 300s  
                             of inactivity.                                     
[01/30/25 06:04:38] INFO     Initialized request thread pool with 8 threads.    
[01/30/25 06:04:38] INFO     Application startup complete.                      
[01/30/25 16:49:36] INFO     Loading visual model 'ViT-B-32__openai' to memory  
[01/30/25 16:49:36] INFO     Setting execution providers to                     
                             ['CPUExecutionProvider'], in descending order of   
                             preference                                         
[01/30/25 16:49:37] INFO     Loading detection model 'buffalo_l' to memory      
[01/30/25 16:49:37] INFO     Setting execution providers to                     
                             ['CPUExecutionProvider'], in descending order of   
                             preference                                         
[01/30/25 16:49:39] INFO     Loading recognition model 'buffalo_l' to memory    
[01/30/25 16:49:39] INFO     Setting execution providers to                     
                             ['CPUExecutionProvider'], in descending order of   
                             preference

Additional information

hwaccel.ml.yaml:

services:
  armnn:
    devices:
      - /dev/mali0:/dev/mali0
    volumes:
      - /lib/firmware/mali_csffw.bin:/lib/firmware/mali_csffw.bin:ro # Mali firmware for your chipset (not always required depending on the driver)
      - /usr/lib/aarch64-linux-gnu/libmali.so:/usr/lib/libmali.so:ro # Mali driver for your chipset (always required)

verifying devices exist on machine:

ls -la /dev/mali0
crw-rw-rw- 1 root root 10, 121 Jan 29 15:59 /dev/mali0

ls -la /lib/firmware/mali_csffw.bin
-rw-r--r-- 1 root root 278528 Jul 28  2020 /lib/firmware/mali_csffw.bin

ls -la /usr/lib/aarch64-linux-gnu/libmali.so
lrwxrwxrwx 1 root root 12 Jul 28  2020 /usr/lib/aarch64-linux-gnu/libmali.so -> libmali.so.1