[Detector Support]: Indexing keeps going

[sharpcall]

Describe the problem you are having

Is indexing supposed to be ongoing always?
This goes to 29% and gets stuck or to 31% and so on. I haven't seen the search so far.

I have google coral.

Version

Installation method Home Assistant OS Core 2025.7.0 Supervisor 2025.06.2 Operating System 14.2 Frontend 20250702.0

Frigate config file

mqtt:
  host: 192.168.2.41 #Insert the IP address of your Home Assistant
  port: 1883 #Leave as default 1883 or change to match the port set in yout MQTT Broker configuration
  user: mqttcamera
  password: kljdJdj3242Jddd
  topic_prefix: frigate
  client_id: frigate
  stats_interval: 60

detectors:
  coral:
    type: edgetpu
    device: usb

database:
  path: /config/frigate.db

# Optional: logger verbosity settings
logger:
  # Optional: Default log verbosity (default: shown below)
  default: info
  # Optional: Component specific logger overrides
  logs:
    frigate.event: debug

go2rtc:
  streams:
    garage:
      - rtsp://admin:Rootme@[email protected]:554/cam/realmonitor?channel=1&subtype=2#video=copy#audio=aac
    front_left_2lens:
      - rtsp://admin:Rootme@[email protected]:554/cam/realmonitor?channel=1&subtype=2#video=copy#audio=aac

cameras:
  garage: #Change this to a name of the camera that makes sense to you
    ffmpeg:
      output_args:
        record: preset-record-generic-audio-aac #Insert this if your camera supports audio output
      inputs:
        - path: rtsp://127.0.0.1:8554/garage #Change as per your camera documentation for high quality feed
          roles:
            - record
        - path: rtsp://127.0.0.1:8554/garage #Change as per your camera documentation for low quality feed
          roles:
            - detect
    detect:
      height: 720 #Change this to match the resolution of your detection channel (in this case channel 1)
      width: 1280 #Change this to match the resolution of your detection channel (in this case channel 1)
      fps: 5 #This is the frame rate for detection, between 5-10 fps is sufficient.
    record:
      enabled: true
      retain:
        days: 3 #The number of days a recording will be kept for after a motion is detected.
        mode: motion
      alerts:
        retain:
          days: 3
      detections:
        retain:
          days: 3
    objects:
      track:
        - person
        - car


version: 0.15-1
camera_groups:
  Exterior:
    order: 1
    icon: LuAlertOctagon
    cameras: birdseye

  Birdseye_View:
    order: 2
    icon: LuBird
    cameras:
      - birdseye
      - reolink_camera
      - garage
semantic_search:
  enabled: true
  reindex: true
  model_size: large

docker-compose file or Docker CLI command

NA

Relevant Frigate log output

2025-07-03 21:41:14.358574645  [2025-07-03 17:41:14] frigate.util.config            INFO    : Checking if frigate config needs migration...
2025-07-03 21:41:14.377476588  [2025-07-03 17:41:14] frigate.util.config            INFO    : frigate config does not need migration...
2025-07-03 21:41:14.441276918  [2025-07-03 17:41:14] frigate.util.services          WARNING : Did not detect hwaccel, using a GPU for accelerated video decoding is highly recommended
2025-07-03 21:41:14.458367924  [2025-07-03 17:41:14] frigate.app                    INFO    : Starting Frigate (0.15.0-6cb5cfb)
2025-07-03 21:41:14.459113525  [2025-07-03 17:41:14] frigate.app                    INFO    : Creating directory: /tmp/cache
2025-07-03 21:41:14.466943364  [2025-07-03 17:41:14] peewee_migrate.logs            INFO    : Starting migrations
2025-07-03 21:41:14.467458822  [2025-07-03 17:41:14] peewee_migrate.logs            INFO    : There is nothing to migrate
2025-07-03 21:41:14.497594523  [2025-07-03 17:41:14] frigate.app                    INFO    : Recording process started: 285
2025-07-03 21:41:14.497598537  [2025-07-03 17:41:14] frigate.app                    INFO    : Review process started: 287
2025-07-03 21:41:14.499346491  [2025-07-03 17:41:14] frigate.app                    INFO    : go2rtc process pid: 114
2025-07-03 21:41:14.518636127  [2025-07-03 17:41:14] detector.coral                 INFO    : Starting detection process: 301
2025-07-03 21:41:14.519131869  [2025-07-03 17:41:14] frigate.app                    INFO    : Embedding process started: 303
2025-07-03 21:41:14.546343152  [2025-07-03 17:41:14] frigate.app                    INFO    : Output process started: 326
2025-07-03 21:41:14.608372902  [2025-07-03 17:41:14] frigate.app                    INFO    : Camera processor started for garage: 340
2025-07-03 21:41:14.608745095  [2025-07-03 17:41:14] frigate.app                    INFO    : Capture process started for garage: 344
2025-07-03 21:41:14.753165707  [2025-07-03 17:41:14] frigate.api.fastapi_app        INFO    : Starting FastAPI app
2025-07-03 21:41:15.054789097  [2025-07-03 17:41:15] frigate.api.fastapi_app        INFO    : FastAPI started
2025-07-03 21:41:16.417915713  [2025-07-03 17:41:16] frigate.util.model             WARNING : OpenVINO failed to build model, using CPU instead: Exception from src/inference/src/cpp/core.cpp:121:
2025-07-03 21:41:16.417922133  Check '!m_device_map.empty()' failed at src/plugins/intel_gpu/src/plugin/plugin.cpp:379:
2025-07-03 21:41:16.417928853  [GPU] Can't get OPTIMIZATION_CAPABILITIES property as no supported devices found or an error happened during devices query.
2025-07-03 21:41:16.417972824  [GPU] Please check OpenVINO documentation for GPU drivers setup guide.
2025-07-03 18:47:50.358010  2025-07-03 21:41:16.417975397
2025-07-03 18:47:50.358033  2025-07-03 21:41:16.417976827
2025-07-03 18:47:50.358037  2025-07-03 21:41:16.417978435
2025-07-03 21:41:17.167462782  [2025-07-03 17:41:17] frigate.embeddings.embeddings  INFO    : Indexing tracked object embeddings...
2025-07-03 21:41:17.281500308  [2025-07-03 17:41:14] frigate.detectors.plugins.edgetpu_tfl INFO    : Attempting to load TPU as usb
2025-07-03 21:41:17.295340026  [2025-07-03 17:41:17] frigate.detectors.plugins.edgetpu_tfl INFO    : TPU found
2025-07-03 21:43:54.723301260  [2025-07-03 17:43:54] frigate.video                  ERROR   : garage: Unable to read frames from ffmpeg process.
2025-07-03 21:43:54.724595915  [2025-07-03 17:43:54] frigate.video                  ERROR   : garage: ffmpeg process is not running. exiting capture thread...
2025-07-03 21:44:04.696136464  [2025-07-03 17:44:04] watchdog.garage                ERROR   : Ffmpeg process crashed unexpectedly for garage.
2025-07-03 21:44:04.696160835  [2025-07-03 17:44:04] watchdog.garage                ERROR   : The following ffmpeg logs include the last 100 lines prior to exit.
2025-07-03 21:44:04.696434848  [2025-07-03 17:44:04] ffmpeg.garage.detect           ERROR   : [swscaler @ 0x7f2b98015340] deprecated pixel format used, make sure you did set range correctly
2025-07-03 21:44:04.696437472  [2025-07-03 17:44:04] ffmpeg.garage.detect           ERROR   : [in#0/rtsp @ 0x562a5c42cd80] Error during demuxing: Connection timed out
2025-07-03 21:44:04.710554716  [2025-07-03 17:44:04] ffmpeg.garage.record           ERROR   : [segment @ 0x55fa43aba240] Timestamps are unset in a packet for stream 0. This is deprecated and will stop working in the future. Fix your code to set the timestamps properly
2025-07-03 21:44:04.710952692  [2025-07-03 17:44:04] ffmpeg.garage.record           ERROR   : [aac @ 0x55fa43a86440] Queue input is backward in time
2025-07-03 21:44:04.710955279  [2025-07-03 17:44:04] ffmpeg.garage.record           ERROR   : [in#0/rtsp @ 0x55fa43a4fe80] Error during demuxing: Connection timed out
2025-07-03 21:44:04.710956777  [2025-07-03 17:44:04] watchdog.garage                INFO    : Terminating the existing ffmpeg process...
2025-07-03 21:44:04.710958122  [2025-07-03 17:44:04] watchdog.garage                INFO    : Waiting for ffmpeg to exit gracefully...
2025-07-03 22:16:14.642177871  [2025-07-03 18:16:14] frigate.storage                INFO    : Less than 1 hour of recording space left, running storage maintenance...
2025-07-03 22:16:14.805261771  [2025-07-03 18:16:14] frigate.storage                INFO    : Cleaned up 1629.1000000000004 MB of recordings

Install method

Home Assistant Add-on

Object Detector

Coral

Screenshots of the Frigate UI's System metrics pages

Any other information that may be helpful

2025-07-03 21:41:10.109607838 [INFO] Preparing new go2rtc config...
2025-07-03 21:41:10.177450560 [INFO] Got IP address from supervisor: 192.168.2.41
2025-07-03 21:41:10.260677301 [INFO] Got WebRTC port from supervisor: 8555
2025-07-03 21:41:11.098002793 [INFO] Starting go2rtc...
2025-07-03 21:41:11.221185962 17:41:11.221 INF go2rtc platform=linux/amd64 revision=b2399f3 version=1.9.2
2025-07-03 21:41:11.221541412 17:41:11.221 INF config path=/dev/shm/go2rtc.yaml
2025-07-03 21:41:11.222826329 17:41:11.222 INF [api] listen addr=:1984
2025-07-03 21:41:11.223159493 17:41:11.222 INF [rtsp] listen addr=:8554
2025-07-03 21:41:11.223699644 17:41:11.223 INF [webrtc] listen addr=:8555/tcp
2025-07-03 21:41:20.127851757 [INFO] Starting go2rtc healthcheck service...
2025-07-03 21:43:49.527067434 17:43:49.526 WRN github.com/AlexxIT/go2rtc/internal/streams/producer.go:171 > error="read tcp 172.30.33.3:40596->192.168.2.115:554: i/o timeout" url=rtsp://admin:Rootme@[email protected]:554/cam/realmonitor?channel=1&subtype=2#video=copy#audio=aac
2025-07-03 22:26:36.968125551 18:26:36.967 WRN github.com/AlexxIT/go2rtc/internal/streams/producer.go:171 > error="RTP header size insufficient: 2 < 4" url=rtsp://admin:Rootme@[email protected]:554/cam/realmonitor?channel=1&subtype=2#video=copy#audio=aac

[hawkeye217]

The error messages in the log are indicating that your GPU failed to work with the large Semantic Search model. You should use the small model instead. There is a negligible difference in quality.

https://docs.frigate.video/configuration/semantic_search/