[Config Support]: additional config needed for macOS M1 docker install/setup?

[ronbruins]

Describe the problem you are having

I have setup my MacBook Air M1 and was able to get frigate up and running. Do I / Can I configure anything that allows me make a better use of the resources? like GPU usage or neural engine?

Version

0.14

Frigate config file

mqtt:
  enabled: true
  host: 192.168.1.35
  port: 1883
  topic_prefix: frigate
  client_id: frigate
  user: **********
  password: **********

# logger:
#   # Optional: default log level (default: shown below)
#   default: debug
tls:
  enabled: false

ui:
  time_format: 24hour
  date_style: short


go2rtc:
  log:
    format: text
    level: error
    #exec: trace
  streams:
    deurbel:
      - rtsp://*****:******@192.168.1.148:554/h264Preview_01_main
      - ffmpeg:achter#audio=opus

detectors:
  cpu1:
    type: cpu
model:
  path: plus://**********
  labelmap:
    16: Kat

snapshots:
  enabled: true
  retain:
    default: 1

detect:
  enabled: true # <---- turn on detection

record:
  enabled: true
  retain:
    days: 0
    mode: all
  events:
    retain:
      default: 7
      mode: motion

objects:
  track:
    - person
    - face
    - license_plate
    - dog
    - cat
    - car
    - amazon
    - fedex
    - ups
    - package
  filters:
    cat:
      min_area: 3000
      threshold: 0.83
    dog:
      min_area: 3000
      threshold: 0.8
    person:
      min_area: 5500
      threshold: 0.8

cameras:
  deurbel:
    ffmpeg:
      inputs:
        - path: rtsp://localhost:8554/deurbel?video=copy&audio=aac
          input_args: preset-rtsp-restream
          roles:
            - record
            - detect
      output_args:
        record: -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c:v copy -c:a aac -tag:v hvc1
    detect:
      width: 1280
      height: 960
    live:
      stream_name: deurbel
    motion:
      threshold: 75
      mask:
        - 0.216,0.548,0.294,0.525,0.296,0.431,0.214,0.43
        - 0.644,0.035,0.492,0.036,0.492,0.002,0.647,0
    objects:
      filters:
        car:
          mask:
            - 0.738,0.689,0.72,0.788,0.141,0.818,0.116,0.736,0.508,0.715
            - 0.736,0.425,0.743,0.623,0,0.642,0,0.476
        cat:
          mask: 0.004,0.005,0.002,0.204,0.234,0.221,0.283,0.008

    zones:
      Muurtje:
        coordinates: 0.124,0.938,0.336,0.999,0.341,0.925,0.149,0.892
        loitering_time: 0
        objects: cat

Relevant Frigate log output

no log

Relevant go2rtc log output

no log

Frigate stats

No response

Operating system

Other

Install method

Docker Compose

docker-compose file or Docker CLI command

services:
  frigate:
    container_name: frigate_014B
    privileged: true # this may not be necessary for all setups
    network_mode: "host"
    restart: unless-stopped
    image: ghcr.io/blakeblackshear/frigate:stable-standard-arm64
    shm_size: "256mb" # update for your cameras based on calculation above

    volumes:
      - /etc/localtime:/etc/localtime:ro
      - /var/root/docker/frigate/config:/config
      - /var/root/docker/frigate/storage:/media/frigate
      - type: tmpfs # Optional: 1GB of memory, reduces SSD/SD Card wear
        target: /tmp/cache
        tmpfs:
          size: 1000000000
      #ports:
      #- "5000:5000"
      #- "1984:1984"
      #- "8554:8554" # RTSP feeds
      #- "8555:8555/tcp" # WebRTC over tcp
      #- "8555:8555/udp" # WebRTC over udp
    environment:
      FRIGATE_RTSP_PASSWORD: "*******"
      PLUS_API_KEY: "************"

Object Detector

CPU (no coral)

Screenshots of the Frigate UI's System metrics pages

No response

Any other information that may be helpful

No response

[NickM-27]

The neural engine is used automatically with the CPU detector. It's impossible to use the GPU because docker on macos does not support device passthrough

[ronbruins]

The neural engine is used automatically with the CPU detector. It's impossible to use the GPU because docker on macos does not support device passthrough

So just the cpu detector should be enough, but do I need add more (cpu 2,3 ..) or will just one use all cores and threads?

[NickM-27]

depends if you have skipped fps, the default should be fine for 6 cameras or so, otherwise can just add another cpu detector

[ronbruins]

Ok and last but not least, I read that the M1 is powerful enough to handle frigate (I have 4 cameras) without using a coral? Would it (could it) make sense to split up, to put go2rtc on one machine(M1) and the frigate on another (Linux)and then using the coral on that one? Or would all on the m1 be fine?

[NickM-27]

all on m1 should be fine