Skip to content
View in the app

A better way to browse. Learn more.

Unraid

A full-screen app on your home screen with push notifications, badges and more.

To install this app on iOS and iPadOS
  1. Tap the Share icon in Safari
  2. Scroll the menu and tap Add to Home Screen.
  3. Tap Add in the top-right corner.
To install this app on Android
  1. Tap the 3-dot menu (⋮) in the top-right corner of the browser.
  2. Tap Add to Home screen or Install app.
  3. Confirm by tapping Install.

[SUPPORT] blakeblackshear - Frigate

Featured Replies

  • Author
17 hours ago, NoRaid99 said:

Does this depend to the developers from frigate or unraid?

I have the strange feeling that this future should be now. :)

At the moment semantic search set to "small model size" is working in a new created container never started with internet connection.

I will keep the setting like this and observe if there are be problems in future.

You can participate in the development https://github.com/blakeblackshear/frigate

There is no "development" in unraid, I just created and maintain the template to run this app in unraid.

  • Replies 1.6k
  • Views 559.4k
  • Created
  • Last Reply

Top Posters In This Topic

Most Popular Posts

  • Glodenox
    Glodenox

    I ran into an issue where the Docker container didn't want to start any more after the upgrade to Unraid 7.3.0, with the following error: Error response from daemon: error while creating mount source

  • I released now a Plugin for Unraid 6.9.0rc2 and up so that the installation of the Drivers is as easy as going to the CA App and search for 'Coral-Driver'.   It will take about an hour to sh

  • Hi @mathgoy, I've been using Frigate with an NVIDIA GPU.   to set it up, go the the docker template and add: --rm --runtime=nvidia to the "Extra Parameters".   Y

Posted Images

  • Author
13 hours ago, theMs said:

I'm having an issue where I click "install" on the docker image, fill out all of the info and then when it's pulling down the images it seems to hang on one part and never completes. Has anyone else run into something like that?

Screenshot 2025-07-15 140326.png

I guess your Docker image is not big enough for the tensorrt image. Go to Settings > Docker, then disable Docker and increase the Docker Image Size to at least 8GB. This tensort version of frigate requires at least 4-5GB.

5 hours ago, yayitazale said:

I guess your Docker image is not big enough for the tensorrt image. Go to Settings > Docker, then disable Docker and increase the Docker Image Size to at least 8GB. This tensort version of frigate requires at least 4-5GB.

I use a disk, not an image so size isn't an issue, it does have plenty of space.

  • Author
16 hours ago, theMs said:

I use a disk, not an image so size isn't an issue, it does have plenty of space.

I don't know what issue can you have then.

  • 2 weeks later...

Hey all,

Testing our Frigate in Unraid to see if it meets my needs and it seems like my NVIDIA 2060 isn't being used for encoding/decoding and I was wondering if someone could offer a hand? Been pouring over docs, posts, etc the last few days so I'm probably a little blind to a resolution at this point.

I'll share some configs below:

image.png


Frigate config:

mqtt:
  enabled: true
  host: 192.168.0.200
  user: '{FRIGATE_MQTT_USER}'
  password: '{FRIGATE_MQTT_PASSWORD}'

detectors:
  coral:
    type: edgetpu
    device: usb

ffmpeg:
  hwaccel_args: preset-nvidia
  output_args:
    record: preset-record-generic-audio-aac

go2rtc:
  streams:
    testcam_fullres:
      - rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0
      - ffmpeg:testcam_fullresolution#audio=aac
    testcam:
      - ffmpeg:rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0#audio=aac#video=h264#encoder=h264_nvenc#width=1280#height=720#bitrate=2500k#fps=15#preset=llhq#ffmpeg=-ar 44100

review:
  alerts:
    labels:
      - person
      - car
  detections:
    labels:
      - dog
      - cat

cameras:
  testcam:
    enabled: true
    ffmpeg:
      inputs:
        - path: rtsp://localhost:8554/testcam_fullres 
          roles:
            - record
        - path: rtsp://localhost:8554/testcam
          roles:
            - detect
    detect:
      enabled: true
      width: 1280
      height: 720
    motion:
      mask: 0.01,0.938,0.009,0.985,0.284,0.987,0.284,0.938
    objects:
      track:
        - person
        - car
        - dog
        - cat

    live:
      stream_name: testcam

record:
  enabled: true
  retain:
    days: 5
    mode: all

  alerts:
    retain:
      days: 30
      mode: motion

  detections:
    retain:
      days: 30
      mode: motion

snapshots:
  enabled: true
  retain:
      default: 30
      objects:
        person: 30
        car: 30
        cat: 30
        dog: 30

version: 0.15-1




2025-07-25 19:54:17.961952591 [INFO] Preparing Frigate...

2025-07-25 19:54:18.040549921 [INFO] Starting Frigate...

2025-07-25 19:54:19.729201612 [2025-07-25 19:54:19] frigate.util.config INFO : Checking if frigate config needs migration...

2025-07-25 19:54:19.742543992 [2025-07-25 19:54:19] frigate.util.config INFO : frigate config does not need migration...

2025-07-25 19:54:19.767363875 [2025-07-25 19:54:19] frigate.app INFO : Starting Frigate (0.15.2-3bda638)

2025-07-25 19:54:19.768138686 [2025-07-25 19:54:19] frigate.util.services INFO : Current file limits - Soft: 40960, Hard: 40960

2025-07-25 19:54:19.768170646 [2025-07-25 19:54:19] frigate.util.services INFO : File limit set. New soft limit: 40960, Hard limit remains: 40960

2025-07-25 19:54:19.775210832 [2025-07-25 19:54:19] peewee_migrate.logs INFO : Starting migrations

2025-07-25 19:54:19.775866648 [2025-07-25 19:54:19] peewee_migrate.logs INFO : There is nothing to migrate

2025-07-25 19:54:19.789804641 [2025-07-25 19:54:19] frigate.app INFO : Recording process started: 382

2025-07-25 19:54:19.800901647 [2025-07-25 19:54:19] frigate.app INFO : Review process started: 391

2025-07-25 19:54:19.802844921 [2025-07-25 19:54:19] frigate.app INFO : go2rtc process pid: 98

2025-07-25 19:54:19.819370203 [2025-07-25 19:54:19] detector.coral INFO : Starting detection process: 398

2025-07-25 19:54:19.836357536 [2025-07-25 19:54:19] frigate.app INFO : Output process started: 417

2025-07-25 19:54:19.864411136 [2025-07-25 19:54:19] frigate.app INFO : Camera processor started for testcam: 431

2025-07-25 19:54:19.885166805 [2025-07-25 19:54:19] frigate.app INFO : Capture process started for testcam: 439

2025-07-25 19:54:20.228884752 [2025-07-25 19:54:20] frigate.api.fastapi_app INFO : Starting FastAPI app

2025-07-25 19:54:20.320991177 [2025-07-25 19:54:20] frigate.api.fastapi_app INFO : FastAPI started

2025-07-25 19:54:22.445554555 [2025-07-25 19:54:19] frigate.detectors.plugins.edgetpu_tfl INFO : Attempting to load TPU as usb

2025-07-25 19:54:22.448760781 [2025-07-25 19:54:22] frigate.detectors.plugins.edgetpu_tfl INFO : TPU found


image.png

image.png

Running nvidia-smi in my unraid dashboard console:

image.png

Running nvidia-smi in my frigate docker console:

image.png

Nvidia Driver Version: 575.64.05

Appreciate any help in advance, ty ty!

Edited by MedievalKnievel
formatting

Howdy,

That’s correct. I use coral for detection as it’s much more efficient than a GPU.

My purposes for the GPU is to encode/decode my primary camera stream down to a desired resolution to serve to home assistant instance.

The out of box substream of my armcrest camera is not to my satisfaction. Thus I resize the stream. This resize operation should be occurring against the GPU which should churn the stream quite easily but as I’ve shown in the images it seems frigate doesn’t leverage the GPU for this encoding/decoding.

I’m guessing it’s something related to my config options but I’ve defined a global FFMPEG value for hardware acceleration to use the GPU.

With this config applied I had assumed the decode/encode operation being done via my go2rtc stream for “testcam” would append the global setting.

Maybe I should try setting hw accel on the go2rtc steam directly? I’d have to check the docs to see if that’s possible but I think all ffmpeg args are supported if I remember correctly

Edited by MedievalKnievel
More clarification

Hi,

okay I see. Sorry that I didn't realise GPU is for resizing the streams only.

I think it shouldn't make any difference if you set FFMPEG global or on a specific source. For me global FFMPEG settings works, but I have no dGPU. ...have a try...

Maybe have a try without go2rtc and set gpu directly in FFMEG. Driver and docker integration settings seem to be correct.

Sorry that I can not help you, have a nice sunday anyway.


...maybe try a older GPU driver? 570...

Edited by NoRaid99

  • Author
22 hours ago, MedievalKnievel said:

Howdy,

That’s correct. I use coral for detection as it’s much more efficient than a GPU.

My purposes for the GPU is to encode/decode my primary camera stream down to a desired resolution to serve to home assistant instance.

The out of box substream of my armcrest camera is not to my satisfaction. Thus I resize the stream. This resize operation should be occurring against the GPU which should churn the stream quite easily but as I’ve shown in the images it seems frigate doesn’t leverage the GPU for this encoding/decoding.

I’m guessing it’s something related to my config options but I’ve defined a global FFMPEG value for hardware acceleration to use the GPU.

With this config applied I had assumed the decode/encode operation being done via my go2rtc stream for “testcam” would append the global setting.

Maybe I should try setting hw accel on the go2rtc steam directly? I’d have to check the docs to see if that’s possible but I think all ffmpeg args are supported if I remember correctly

You don't need to use "--gpus all" as the GPU pass-through of a specific GPU to the container is done with the "nvidia visible devices" and "nvidia driver capabilities" variable. If you are using the GPU only for decoding/encoding then you can delete the YOLO_MODELS variable as this is only use for GPU detection.

What you can see about the GPU decoding in the logs of frigate? Is go2rtc working correctly?

Edited by yayitazale

2 hours ago, yayitazale said:

You don't need to use "--gpus all" as the GPU pass-through of a specific GPU to the container is done with the "nvidia visible devices" and "nvidia driver capabilities" variable. If you are using the GPU only for decoding/encoding then you can delete the YOLO_MODELS variable as this is only use for GPU detection.

What you can see about the GPU decoding in the logs of frigate? Is go2rtc working correctly?


SOLVED!!!!


I'll leave the additional details below in case they help anyone in the future but the ticket was adding #hardware to my resize ffmpeg params for the testcam stream:


go2rtc:
  streams:
    testcam_fullres:
      - rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0
      - ffmpeg:testcam_fullresolution#audio=aac
    testcam:
      - ffmpeg:rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0#audio=aac#video=h264#encoder=h264_nvenc#width=1280#height=720#bitrate=2500k#fps=15#preset=llhq#hardware




EDIT: please feel free to review the below, but I modified my go2rtc config and theres some progress. Previously my configuration was:

go2rtc:
  streams:
    testcam_fullres:
      - rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0
      - ffmpeg:testcam_fullresolution#audio=aac
    testcam:
      - ffmpeg:rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0#audio=aac#video=h264#encoder=h264_nvenc#width=1280#height=720#bitrate=2500k#fps=15#preset=llhq#ffmpeg=-ar 44100

The above results in a 720p stream to HA, I then modified my go2rtc to:

go2rtc:
  streams:
    testcam_fullres:
      - rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0
      - ffmpeg:testcam_fullresolution#audio=aac
    testcam:
      - rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0
      - "ffmpeg:testcam#audio=aac#video=h264#encoder=h264_nvenc#width=1280#height=720#bitrate=2500k"

I can now see my GPU being used for decoding, but the stream to HA is back to the fullresolution stream from the camer, it seems the ffmpeg args aren't being applied to the stream? From VLC to rtsp://192.168.0.200:8554/testcam:


image.png

So I'm wondering if this is just some misconfiguration with my go2rtc and ffmpeg args?
image.png

Edit 2:

Checking go2rtc via port 1984, I see three streams:

image.png

Checking the producer details of camera.testcam:

image.png

So the params seem to be on the URL? But theyre not transforming the stream?

Edit 3:

Reverting my config to 720 via:


go2rtc:
  streams:
    testcam_fullres:
      - rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0
      - ffmpeg:testcam_fullresolution#audio=aac
    testcam:
      - ffmpeg:rtsp://{FRIGATE_RTSP_USER}:{FRIGATE_RTSP_PASSWORD}@192.168.50.2:554/cam/realmonitor?channel=1&subtype=0#audio=aac#video=h264#encoder=h264_nvenc#width=1280#height=720#bitrate=2500k#fps=15#preset=llhq#ffmpeg=-ar 44100

Results in 2-stream in rtc

image.png

testcam (720p) shows:

image.png



Hello,


Thanks for taking a look!

I've removed the YOLO_MODELS as well as the --gpus all extra parameter and it appears to be the same result - 0 gpu activity for encoding/decoding.

Regarding logging, this is all the information available to me in logs via Unraid, not sure why that is. As I interact with Frigate, via HA for example, I do see API request get logged, but not much else after startup, would you have any recommendations regarding additional logging? The Frigate UI returns the same logging statements:


s6-rc: info: service nginx successfully started

s6-rc: info: service certsync: starting

s6-rc: info: service certsync successfully started

s6-rc: info: service legacy-services: starting

s6-rc: info: service legacy-services successfully started

2025-07-27 17:16:16.984157544 [INFO] Starting go2rtc...

2025-07-27 17:16:17.067737312 17:16:17.067 INF go2rtc platform=linux/amd64 revision=b2399f3 version=1.9.2

2025-07-27 17:16:17.067740898 17:16:17.067 INF config path=/dev/shm/go2rtc.yaml

2025-07-27 17:16:17.068138729 17:16:17.068 INF [api] listen addr=:1984

2025-07-27 17:16:17.068204363 17:16:17.068 INF [rtsp] listen addr=:8554

2025-07-27 17:16:17.068312326 17:16:17.068 INF [webrtc] listen addr=:8555/tcp

2025-07-27 17:16:17.575658969 [INFO] Starting certsync...

2025-07-27 17:16:18.855746669 [2025-07-27 17:16:18] frigate.util.config INFO : Checking if frigate config needs migration...

2025-07-27 17:16:18.868999035 [2025-07-27 17:16:18] frigate.util.config INFO : frigate config does not need migration...

2025-07-27 17:16:18.895554684 [2025-07-27 17:16:18] frigate.app INFO : Starting Frigate (0.15.2-3bda638)

2025-07-27 17:16:18.896346798 [2025-07-27 17:16:18] frigate.util.services INFO : Current file limits - Soft: 40960, Hard: 40960

2025-07-27 17:16:18.896374571 [2025-07-27 17:16:18] frigate.util.services INFO : File limit set. New soft limit: 40960, Hard limit remains: 40960

2025-07-27 17:16:18.903793033 [2025-07-27 17:16:18] peewee_migrate.logs INFO : Starting migrations

2025-07-27 17:16:18.904380501 [2025-07-27 17:16:18] peewee_migrate.logs INFO : There is nothing to migrate

2025-07-27 17:16:18.916107481 [2025-07-27 17:16:18] frigate.app INFO : Recording process started: 382

2025-07-27 17:16:18.925019200 [2025-07-27 17:16:18] frigate.app INFO : Review process started: 391

2025-07-27 17:16:18.926952857 [2025-07-27 17:16:18] frigate.app INFO : go2rtc process pid: 98

2025-07-27 17:16:18.941766278 [2025-07-27 17:16:18] detector.coral INFO : Starting detection process: 398

2025-07-27 17:16:18.956788122 [2025-07-27 17:16:18] frigate.app INFO : Output process started: 417

2025-07-27 17:16:18.980374160 [2025-07-27 17:16:18] frigate.app INFO : Camera processor started for testcam: 431

2025-07-27 17:16:18.995085248 [2025-07-27 17:16:18] frigate.app INFO : Capture process started for testcam: 437

2025-07-27 17:16:19.101550732 [2025-07-27 17:16:19] frigate.api.fastapi_app INFO : Starting FastAPI app

2025-07-27 17:16:19.192482597 [2025-07-27 17:16:19] frigate.api.fastapi_app INFO : FastAPI started

2025-07-27 17:16:21.565586554 [2025-07-27 17:16:18] frigate.detectors.plugins.edgetpu_tfl INFO : Attempting to load TPU as usb

2025-07-27 17:16:21.568974113 [2025-07-27 17:16:21] frigate.detectors.plugins.edgetpu_tfl INFO : TPU found

2025-07-27 17:16:26.311426543 [INFO] Starting go2rtc healthcheck service...

Within Frigate if i access logs > Go2rtc I see:

2025-07-27 17:16:16.308238201 [INFO] Preparing new go2rtc config...

2025-07-27 17:16:16.984157544 [INFO] Starting go2rtc...

2025-07-27 17:16:17.067737312 17:16:17.067 INF go2rtc platform=linux/amd64 revision=b2399f3 version=1.9.2

2025-07-27 17:16:17.067740898 17:16:17.067 INF config path=/dev/shm/go2rtc.yaml

2025-07-27 17:16:17.068138729 17:16:17.068 INF [api] listen addr=:1984

2025-07-27 17:16:17.068204363 17:16:17.068 INF [rtsp] listen addr=:8554

2025-07-27 17:16:17.068312326 17:16:17.068 INF [webrtc] listen addr=:8555/tcp

2025-07-27 17:16:26.311426543 [INFO] Starting go2rtc healthcheck service...

2025-07-27 17:33:29.942486750 17:33:29.942 INF [streams] create new stream url=rtsp://192.168.0.200:8554/testcam

2025-07-27 17:33:29.944609504 17:33:29.944 INF [streams] create new stream url=rtsp://192.168.0.200:8554/testcam

2025-07-27 17:33:41.904474459 17:33:41.904 INF [streams] create new stream url=rtsp://192.168.0.200:8554/testcam

2025-07-27 17:33:41.905668341 17:33:41.905 INF [streams] create new stream url=rtsp://192.168.0.200:8554/testcam

From within the Frigate metrics view:

image.png

And from within Unraid dashboard:

image.png

So it definitely appears that the GPU is passed through to Frigate, and Frigate is connected but it just doesn't use the GPU for encode/decode which is why I leaned back towards an error in configuration.

I would echo your perspective and was hoping to see SOMETHING more in logs, the lack of information is definitely problematic, not sure why logs are so quiet

Edited by MedievalKnievel

  • 3 weeks later...

I've been trying to set up Frigate with OpenVINO on my Unraid server and have been running into persistent stability issues with iGPU hardware acceleration. I'm hoping to share my findings in case it helps developers or other users.

System Details:

  • Unraid Version: 7.0.1

  • CPU: Intel® Core i7-7700 with HD Graphics 630

  • Docker: Frigate (latest stable) with OpenVINO detector

Problem Description:
The Frigate container starts fully with iGPU acceleration enabled, and I can even see the live camera feed for a short period (a few seconds). However, the container consistently crashes and restarts. The primary error in the Frigate logs is:

frigate.watchdog INFO : Detection appears to have stopped. Exiting Frigate...

This indicates that the OpenVINO detection process, running on the iGPU, is freezing or becoming unresponsive.

Troubleshooting Steps Taken:
I've followed all the standard recommendations to stabilize Intel iGPU passthrough on Unraid:

  1. Syslinux/Boot Parameters: Added intel_iommu=on iommu=pt i915.enable_guc=2 to the append line.

  2. Docker Container Settings:

    • Set to Privileged: ON.

    • Passed the device using Extra Parameters: --device=/dev/dri:/dev/dri.

    • Increased shared memory with mount -o remount,size=1G /dev/shm.

  3. Frigate Config (config.yml):

    • Used preset-vaapi for hwaccel_args.

    • Configured the OpenVINO detector with device: iGPU.

Despite all these measures, the setup remains unstable and crashes after a few minutes of operation.

Working Solution (Workaround):
The system is 100% stable when running Frigate in CPU-only mode:

  • Docker container with Privileged: OFF and no extra device passthrough.

  • Frigate config using type: cpu for the detector and no hwaccel_args.

This confirms the issue is specifically related to the stability of the iGPU driver stack (i915/VAAPI/OpenVINO) within the Docker environment on Unraid. It seems like there might be a deeper conflict or a driver issue with this specific hardware/software combination.

Has anyone else with a Kaby Lake (7th gen) CPU experienced similar iGPU freezes with hardware acceleration?

I now have Unraid 7.1.4. I couldn't get it to work, even though I spent several hours trying. I went back to the CPU and it works. I'll probably try again with the igpu.

  • Author
8 hours ago, Xwint5 said:

I've been trying to set up Frigate with OpenVINO on my Unraid server and have been running into persistent stability issues with iGPU hardware acceleration. I'm hoping to share my findings in case it helps developers or other users.

System Details:

  • Unraid Version: 7.0.1

  • CPU: Intel® Core i7-7700 with HD Graphics 630

  • Docker: Frigate (latest stable) with OpenVINO detector

Problem Description:
The Frigate container starts fully with iGPU acceleration enabled, and I can even see the live camera feed for a short period (a few seconds). However, the container consistently crashes and restarts. The primary error in the Frigate logs is:

frigate.watchdog INFO : Detection appears to have stopped. Exiting Frigate...

This indicates that the OpenVINO detection process, running on the iGPU, is freezing or becoming unresponsive.

Troubleshooting Steps Taken:
I've followed all the standard recommendations to stabilize Intel iGPU passthrough on Unraid:

  1. Syslinux/Boot Parameters: Added intel_iommu=on iommu=pt i915.enable_guc=2 to the append line.

  2. Docker Container Settings:

    • Set to Privileged: ON.

    • Passed the device using Extra Parameters: --device=/dev/dri:/dev/dri.

    • Increased shared memory with mount -o remount,size=1G /dev/shm.

  3. Frigate Config (config.yml):

    • Used preset-vaapi for hwaccel_args.

    • Configured the OpenVINO detector with device: iGPU.

Despite all these measures, the setup remains unstable and crashes after a few minutes of operation.

Working Solution (Workaround):
The system is 100% stable when running Frigate in CPU-only mode:

  • Docker container with Privileged: OFF and no extra device passthrough.

  • Frigate config using type: cpu for the detector and no hwaccel_args.

This confirms the issue is specifically related to the stability of the iGPU driver stack (i915/VAAPI/OpenVINO) within the Docker environment on Unraid. It seems like there might be a deeper conflict or a driver issue with this specific hardware/software combination.

Has anyone else with a Kaby Lake (7th gen) CPU experienced similar iGPU freezes with hardware acceleration?

I now have Unraid 7.1.4. I couldn't get it to work, even though I spent several hours trying. I went back to the CPU and it works. I'll probably try again with the igpu.

@ich777 I don't know if you can help us with this.

9 hours ago, Xwint5 said:

Unraid Version: 7.0.1

I would recommend that you upgrade to 7.1.4 to begin with, never mind I just read that you are already on 7.1.4.

9 hours ago, Xwint5 said:

i915.enable_guc=2

I think that won't change much but it shouldn't hurt to enable GuC and HuC

9 hours ago, Xwint5 said:

intel_iommu=on iommu=pt

Since we are dealing with Docker that won't do much I think since these are primarily for VMs.

9 hours ago, Xwint5 said:

Set to Privileged: ON.

This is not a troubleshooting step since you give the container full access to the host and this is basically bad, Privileged should not be turned on whatsoever except you want to make everything insecure. :)

9 hours ago, Xwint5 said:

Passed the device using Extra Parameters: --device=/dev/dri:/dev/dri.

If your've already have a Device mapping in the template that won't do much, which I assume you had because otherwise that wouldn't work, what basically means this is redundant and not necessary.

9 hours ago, Xwint5 said:

Increased shared memory with mount -o remount,size=1G /dev/shm.

I'm not too familiar with Firgate but you shouldn't have to do that and I'm also not sure if Firgate mounts anything in /dev/shm.

9 hours ago, Xwint5 said:

Frigate Config (config.yml):

  • Used preset-vaapi for hwaccel_args.

  • Configured the OpenVINO detector with device: iGPU.

Can't say anything to that because I'm not running Frigate.

9 hours ago, Xwint5 said:

This confirms the issue is specifically related to the stability of the iGPU driver stack (i915/VAAPI/OpenVINO) within the Docker environment on Unraid. It seems like there might be a deeper conflict or a driver issue with this specific hardware/software combination.

I don't think so because OpenVINO should work no matter what, except if you completely overload it or we are dealing with a hardware failure, or it could even be the case that the your iGPU doesn't fully support the OpenVINO implementation from Frigate <- but I don't think that's the issue.

However it always depends on the implementation from Frigate how OpenVINO is used and you might need to pass through /dev/usb to your container, so to speaking a Device with the mapping: /dev/bus/usb:/dev/bus/usb

Hello,

I'm having trouble with this container and downloading/installing object detection models. I have a model_cache directory within the config folder. In the container; I have defined yolov7-640 as the model I want.

My understand is frigate will automatically download the model but in my case I don't see a yolov7 directory being created in the model_cache directory.

Any guides or help appreciated.

No, Frigate will not automatically download the model, but it has some in the container. I don't recommend trying yolo, because I had a lot of trouble with it, specifically with YOLOv8 (and a few others). Including the u8 != f32 error, I also had to do a lot of preprocessing here. My big problem was a non-standard directory for Config Path, which caused a lot of confusion in Frigate, although it shouldn't have. Currently, igpu works for me on the ssdlite_mobilenet_v2 model (FP16 precision IR model). This is a model from the Frigate documentation. I recommend testing on this model, I currently have Frigate 0.16-0. You also need to do the config. correctly and specify the path to the model /openvino-model/ssdlite_mobilenet_v2.xml. This model is physically located in the docker image layers.

https://docs.frigate.video/configuration/object_detectors/#supported-models-2

Edited by Xwint5

  • Author
On 8/18/2025 at 4:58 PM, repomanz said:

Hello,

I'm having trouble with this container and downloading/installing object detection models. I have a model_cache directory within the config folder. In the container; I have defined yolov7-640 as the model I want.

My understand is frigate will automatically download the model but in my case I don't see a yolov7 directory being created in the model_cache directory.

Any guides or help appreciated.

I don't know how you are launching the container but it is already prepared for yolo models. Are you using the tensor branch? did you have the nvidia driver installed? https://docs.frigate.video/configuration/object_detectors#generate-models

  • 2 weeks later...

I installed the tensor version, but it seems tensor is no longer supported in version 16. So I installed the normal stable release (through the unraid apps section) I also don't see a models folder in my appdata folder on the frigate appdata folder. Can I create those or are they supposed to be created automatically?

  • Author
8 hours ago, KittenMittons said:

I installed the tensor version, but it seems tensor is no longer supported in version 16. So I installed the normal stable release (through the unraid apps section) I also don't see a models folder in my appdata folder on the frigate appdata folder. Can I create those or are they supposed to be created automatically?

Tensor version is supported in version 16 https://github.com/blakeblackshear/frigate/releases/tag/v0.16.0.

Models will be created automatically with tensorrt version if you have correctly defined YOLO_MODELS varibale and the rest of the template https://docs.frigate.video/configuration/object_detectors#generate-models

19 hours ago, yayitazale said:

Tensor version is supported in version 16 https://github.com/blakeblackshear/frigate/releases/tag/v0.16.0.

Models will be created automatically with tensorrt version if you have correctly defined YOLO_MODELS varibale and the rest of the template https://docs.frigate.video/configuration/object_detectors#generate-models

okay so I re-installed the tensor version. I am able to get it up and running and add a camera. It it showing the Nvidia GPU as passed through. This is the yolo version I have set in the template (however I've tried multiple versions and it never generates the models):

image.png

However, it's not loading any of the models in the appdata folder:

image.png

and if I try to add the tensorrt detector to my config, frigate just bootloops until I remove it

detectors:
tensorrt:
type: tensorrt
device: 0

do I need to manually create those model folders/files?

Edited by KittenMittons

  • Author
On 9/4/2025 at 3:42 AM, KittenMittons said:

okay so I re-installed the tensor version. I am able to get it up and running and add a camera. It it showing the Nvidia GPU as passed through. This is the yolo version I have set in the template (however I've tried multiple versions and it never generates the models):

image.png

However, it's not loading any of the models in the appdata folder:

image.png

and if I try to add the tensorrt detector to my config, frigate just bootloops until I remove it

detectors:
tensorrt:
type: tensorrt
device: 0

do I need to manually create those model folders/files?

Did you added "--runtime=nvidia" as extra parameter under advanced view?

5 hours ago, yayitazale said:

Did you added "--runtime=nvidia" as extra parameter under advanced view?

Yessir

  • Author
22 minutes ago, KittenMittons said:

Yessir

And what do you see in the logs?

41 minutes ago, yayitazale said:

And what do you see in the logs?

attached a copy of the log after restarting the container in unraid

frigate-logs.txt

Unable to get GPU detector working on v16.

I just upgrade to v16 and I can no longer get my NVIDIA GPU to work as a detector. I tried keeping the tensorrt settings, but that just throws an error stating it's no longer support and causes the docker container to endlessly restart. I also tried changing it to onnx, but that also causes it to endlessly loop, and I have no idea how to get any models to work with it. The only way for me to keep the container up and running is by using a CPU detector, so I know the rest of the config should be good.

GPU: Nvidia RTX 2060

Docker Image: ghcr.io/blakeblackshear/frigate:stable-tensorrt

YOLO_MODELS: yolov7-320,yolov7x-640

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

Account

Navigation

Search

Search

Configure browser push notifications

Chrome (Android)
  1. Tap the lock icon next to the address bar.
  2. Tap Permissions → Notifications.
  3. Adjust your preference.
Chrome (Desktop)
  1. Click the padlock icon in the address bar.
  2. Select Site settings.
  3. Find Notifications and adjust your preference.