[SUPPORT] blakeblackshear - Frigate


Recommended Posts

5 minutes ago, yayitazale said:

Check the changelog breaking change part and if there is anything in your config that must be changed, read the docs to understand how to change it. Then update and check the logs. If there is no problem, then you can also do the same procedure but for the new features that you can add.

 Yes did that already :) I ment, that I do not have any function on Unraid to update Frigata. Unraid does not offer me to upgrade Frigata container. So the question was how to start the upgrade. 
Or I just don't see the update button 🤷‍♂️

Link to comment
1 minute ago, kpcz said:

 Yes did that already :) I ment, that I do not have any function on Unraid to update Frigata. Unraid does not offer me to upgrade Frigata container. So the question was how to start the upgrade. 
Or I just don't see the update button 🤷‍♂️

There is a quite big button in the docker tab, below all the containers with the name "CHECK FOR UPDATES", if there is any update of any image of any docker, it will appear a message in that container "update". You can also hit "UPDATE ALL" button close to "CHECK FOR UPDATES" once it finished looking for updates of the containers. If you dind't change anything in the community app settings, the update search must work automatically too.

Link to comment
4 minutes ago, yayitazale said:

There is a quite big button in the docker tab, below all the containers with the name "CHECK FOR UPDATES", if there is any update of any image of any docker, it will appear a message in that container "update". You can also hit "UPDATE ALL" button close to "CHECK FOR UPDATES" once it finished looking for updates of the containers. If you dind't change anything in the community app settings, the update search must work automatically too.

I think the problem is that my Frigate container version shows: not available.  (other containers have version number)  So it cannot check for  updates. looks like the only option is remove and reinstall Frigate then. 

I am running unraid 6.11. looks this is fixed in unraid 6.12

Edited by kpcz
Link to comment
1 minute ago, kpcz said:

I think the problem is that my Frigate container version shows: not available.  (other containers have version number)  So it cannot check for  updates. looks like the only option is remove and reinstall Frigate then. 

This sometimes happens in unraid...

  • Like 1
Link to comment

Having some issues getting object detection to work. Recordings/go2rtc/HA integration working but can't get any events to generate.

 

mqtt:
  host: 10.0.0.3
  port: 1883
  topic_prefix: frigate
  client_id: frigate
  user: frigate
  password: **REDACTED**
  stats_interval: 60

ffmpeg:
  hwaccel_args: preset-nvidia-h264

detectors:
  cuda:
    type: tensorrt

model:
  path: /config/model_cache/tensorrt/yolov4-tiny-416.trt
  labelmap_path: /labelmap.txt
  width: 416
  height: 416

record:
  enabled: true
  retain:
    days: 7
    mode: motion
  events:
    retain:
      default: 5
      mode: active_objects
    pre_capture: 3
    post_capture: 10

snapshots:
  enabled: true
  retain:
    default: 30
  timestamp: false
  bounding_box: true

birdseye:
  enabled: false
  restream: false
  width: 1280
  height: 720
  quality: 1
  mode: objects

go2rtc:
  streams:
    driveway:
    - rtsp://frigate:**REDACTED**@10.0.0.8:554/cam/realmonitor?channel=1&subtype=0
    driveway_sub:
    - rtsp://frigate:**REDACTED**@10.0.0.8:554/cam/realmonitor?channel=1&subtype=1
    front:
    - rtsp://frigate:**REDACTED**@10.0.0.6:554/cam/realmonitor?channel=1&subtype=0
    front_sub:
    - rtsp://frigate:**REDACTED**@10.0.0.6:554/cam/realmonitor?channel=1&subtype=1
    front_door:
    - rtsp://frigate:**REDACTED**@10.0.0.5:554/cam/realmonitor?channel=2&subtype=0
    front_door_sub:
    - rtsp://frigate:**REDACTED**@10.0.0.5:554/cam/realmonitor?channel=2&subtype=1
    garage:
    - rtsp://frigate:**REDACTED**@10.0.0.7:554/cam/realmonitor?channel=1&subtype=0
    garage_sub:
    - rtsp://frigate:**REDACTED**@10.0.0.7:554/cam/realmonitor?channel=1&subtype=1
    livingroom:
    - rtsp://homeassistant:**REDACTED**@10.0.0.9:554/cam/realmonitor?channel=1&subtype=0
    livingroom_sub:
    - rtsp://homeassistant:**REDACTED**@10.0.0.9:554/cam/realmonitor?channel=1&subtype=1
    north_deck:
    - rtsp://frigate:**REDACTED**@10.0.0.5:554/cam/realmonitor?channel=4&subtype=0
    north_deck_sub:
    - rtsp://frigate:**REDACTED**@10.0.0.5:554/cam/realmonitor?channel=4&subtype=1
    south_deck:
    - rtsp://frigate:**REDACTED**@10.0.0.5:554/cam/realmonitor?channel=3&subtype=0
    south_deck_sub:
    - rtsp://frigate:**REDACTED**@10.0.0.5:554/cam/realmonitor?channel=3&subtype=1

cameras:
  front_door:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/front_door_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/front_door
        roles:
          - record
    motion:
      mask:
      - 187,7,187,25,7,23,7,7
    zones:
      front_porch:
        coordinates: 105,240,352,240,352,93,275,113,224,69,192,22,126,0
      front_lawn:
        coordinates: 219,67,271,104,298,0,185,0

  north_deck:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/north_deck_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/north_deck
        roles:
        - record
    motion:
      mask:
      - 331,7,330,23,149,25,149,8

  south_deck:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/south_deck_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/south_deck
        roles:
        - record
    motion:
      mask:
      - 333,6,331,24,149,25,151,6

  front:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/front_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/front
        roles:
        - audio
        - record
    motion:
      mask:
      - 331,216,333,238,156,239,158,216
    zones:
      front_lawn:
        coordinates: 240,185,277,240,352,240,352,115,267,79,182,50,146,58,107,69,95,101
      front_porch:
        coordinates: 0,134,89,84,66,43,0,73
      driveway_top:
        coordinates: 63,78,149,56,99,28,53,30

  garage:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/garage_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/garage
        roles:
        - audio
        - record
    objects:
      track:
      - person
      - bicycle
      - car
      - cat
      - dog
    motion:
      mask:
      - 180,0,179,21,0,21,0,0
    zones:
      garage_floor:
        coordinates: 40,164,52,202,123,240,352,240,352,98,274,30,104,27
        objects:
        - person
        - bicycle
        - car
        - cat
        - dog

  driveway:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/driveway_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/driveway
        roles:
        - audio
        - record
    motion:
      mask:
      - 337,213,332,233,159,233,159,214

  livingroom:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/livingroom_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/livingroom
        roles:
        - record
    objects:
      track:
      - person
      - dog
    motion:
      mask:
      - 618,18,619,46,395,45,393,21

 

I also purchased a USB Coral device but have been unable to get that thing to detect. Device seems extremely problematic from what I've gathered from all the issues I've dug through. Without being able to compare performance to my my Nvidia card I'm not sure there's much to gain from getting it to work anyway.


Does anyone see any obvious issues with my config?

Edited by BrandonG777
Stumped
Link to comment

Some of your cameras don't have any objects to track. 

 

I'd start with cpu detector first while you dial it in then swap to coral or nvidia for detection.

 

Note the initial motion detection is done on CPU it then passes the still image to the coral or nvidia for object matching.

 

This means you often need to mask out, motion masks & object masks for areas with lots of excess movement like a tree or large bush in the garden. 

 

So start by commenting # out nvidia detection & comment # out the model. See if you get any object detections & events. Use the camera debug feed. Then start adding back in the nvidia stuff & custom models.

Edited by dopeytree
  • Upvote 1
Link to comment

Thank you. Not exactly sure what my issue was but I believe you were onto something when you talked about objects to track but I kind of took the shotgun approach and just changing a bunch of stuff. Working well now, just need to finish up masking.

Will circle back to the USB Coral device but wish I had just gone PCIe.

Link to comment
6 hours ago, BrandonG777 said:

Thank you. Not exactly sure what my issue was but I believe you were onto something when you talked about objects to track but I kind of took the shotgun approach and just changing a bunch of stuff. Working well now, just need to finish up masking.

Will circle back to the USB Coral device but wish I had just gone PCIe.

I have a USB coral since since more than 4 years ago and I'm so happy with it. You have to know than some coral devices are defective from the factory, but this applies to both USB and PCIe versions.

 

We will help you to find out if this is your case.

Link to comment

Is there any reason there is no pre configured variable for usb passthru? Or is that not needed? 

 

Or there is no mention in this thread, nor in the coral driver plugin, how to do that?

 

shouldnt something liek this be direct in the template and or first page?

 

Only mention i found is in the docker description i should give frigate admin, but cant the usb just be passed thru without higher privileges?

Edited by NewDisplayName
Link to comment
I have a USB coral since since more than 4 years ago and I'm so happy with it. You have to know than some coral devices are defective from the factory, but this applies to both USB and PCIe versions.
 
We will help you to find out if this is your case.

Okay, you’ve motivated me to see this Coral thing through. I read that a lot of the included cables are bad, doesn’t supply enough voltage or something. I did try another sketch cable and got the same result but I’m not convinced either of these cables are doing what they need to. Will try to grab another cable while I’m out today.

One thing I’m hoping for is that the Coral generates a lot less heat than my Nvidia card and leave it freed up for video transcoding. Seems like it will fit the bill.

The Coral device I have is identifying itself with the changing identifier. I believe this has something to do with my issues of it not working? Perhaps you could recommend a procedure for verifying my Coral is actually good?

Anyway, I appreciate the help, thank you, thank you!


Sent from my iPhone using Tapatalk
Link to comment
7 hours ago, BrandonG777 said:


Okay, you’ve motivated me to see this Coral thing through. I read that a lot of the included cables are bad, doesn’t supply enough voltage or something. I did try another sketch cable and got the same result but I’m not convinced either of these cables are doing what they need to. Will try to grab another cable while I’m out today.

One thing I’m hoping for is that the Coral generates a lot less heat than my Nvidia card and leave it freed up for video transcoding. Seems like it will fit the bill.

The Coral device I have is identifying itself with the changing identifier. I believe this has something to do with my issues of it not working? Perhaps you could recommend a procedure for verifying my Coral is actually good?

Anyway, I appreciate the help, thank you, thank you!


Sent from my iPhone using Tapatalk

Did u grant the docker higher rights? That was the only thing, besides config, for me, i got it today running...

Link to comment
3 hours ago, NewDisplayName said:

Yes. I just enabled it and restartet. What is your config?

image.thumb.png.b95a7344f54d83ad17324aa33e64b13d.png

docker run
  -d
  --name='frigate'
  --net='proxynet'
  --privileged=true
  -e TZ="America/Chicago"
  -e HOST_OS="Unraid"
  -e HOST_HOSTNAME="z"
  -e HOST_CONTAINERNAME="frigate"
  -e 'FRIGATE_RTSP_PASSWORD'='**REMOVED**'
  -e 'PLUS_API_KEY'='**REMOVED**'
  -e 'NVIDIA_VISIBLE_DEVICES'='GPU-82933e84-9c02-0ae5-9cf9-bcdd8f64f15d'
  -e 'NVIDIA_DRIVER_CAPABILITIES'='compute,utility,video'
  -e 'YOLO_MODELS'='yolov4-416,yolov4-tiny-416'
  -e 'USE_FP16'='true'
  -e 'TRT_MODEL_PREP_DEVICE'='0'
  -l net.unraid.docker.managed=dockerman
  -l net.unraid.docker.webui='http://[IP]:[PORT:5000]'
  -l net.unraid.docker.icon='https://raw.githubusercontent.com/yayitazale/unraid-templates/main/frigate.png'
  -p '5000:5000/tcp'
  -p '8554:8554/tcp'
  -p '1984:1984/tcp'
  -p '8555:8555/tcp'
  -p '8555:8555/udp'
  -v '/mnt/user/appdata/frigate':'/config':'rw'
  -v '/mnt/user/media/frigate':'/media/frigate':'rw'
  -v '/etc/localtime':'/etc/localtime':'rw'
  --device='/dev/bus/usb'
  --runtime=nvidia
  --shm-size=256mb
  --mount type=tmpfs,target=/tmp/cache,tmpfs-size=1000000000
  --restart unless-stopped 'ghcr.io/blakeblackshear/frigate:stable-tensorrt'
5412a9d43979593111369496272105f33f70df7f5b0a878569905a2173ef9795


 

mqtt:
  host: 10.0.0.3
  port: 1883
  topic_prefix: frigate
  client_id: frigate
  user: frigate
  password: **REMOVED**
  stats_interval: 60

ffmpeg:
  hwaccel_args: preset-nvidia-h264

detectors:
  coral:
    type: edgetpu
    device: usb
  cuda:
    type: tensorrt

model:
  path: /config/model_cache/tensorrt/yolov4-tiny-416.trt
  labelmap_path: /labelmap.txt
  input_tensor: nchw
  width: 416
  height: 416

record:
  enabled: true
  retain:
    days: 7
    mode: motion
  events:
    retain:
      default: 5
      mode: active_objects
    pre_capture: 3
    post_capture: 10

snapshots:
  enabled: true
  retain:
    default: 30
  timestamp: false
  bounding_box: true

birdseye:
  enabled: false
  restream: true
  width: 1280
  height: 720
  quality: 8
  mode: motion

go2rtc:
  streams:
    driveway:
    - rtsp://frigate:**REMOVED**@10.0.0.8:554/cam/realmonitor?channel=1&subtype=0
    driveway_sub:
    - rtsp://frigate:**REMOVED**@10.0.0.8:554/cam/realmonitor?channel=1&subtype=1
    front:
    - rtsp://frigate:**REMOVED**@10.0.0.6:554/cam/realmonitor?channel=1&subtype=0
    front_sub:
    - rtsp://frigate:**REMOVED**@10.0.0.6:554/cam/realmonitor?channel=1&subtype=1
    front_door:
    - rtsp://frigate:**REMOVED**@10.0.0.5:554/cam/realmonitor?channel=2&subtype=0
    front_door_sub:
    - rtsp://frigate:**REMOVED**@10.0.0.5:554/cam/realmonitor?channel=2&subtype=1
    garage:
    - rtsp://frigate:**REMOVED**@10.0.0.7:554/cam/realmonitor?channel=1&subtype=0
    garage_sub:
    - rtsp://frigate:**REMOVED**@10.0.0.7:554/cam/realmonitor?channel=1&subtype=1
    livingroom:
    - rtsp://homeassistant:**REMOVED**@10.0.0.9:554/cam/realmonitor?channel=1&subtype=0
    livingroom_sub:
    - rtsp://homeassistant:**REMOVED**@10.0.0.9:554/cam/realmonitor?channel=1&subtype=1
    north_deck:
    - rtsp://frigate:**REMOVED**@10.0.0.5:554/cam/realmonitor?channel=4&subtype=0
    north_deck_sub:
    - rtsp://frigate:**REMOVED**@10.0.0.5:554/cam/realmonitor?channel=4&subtype=1
    south_deck:
    - rtsp://frigate:**REMOVED**@10.0.0.5:554/cam/realmonitor?channel=3&subtype=0
    south_deck_sub:
    - rtsp://frigate:**REMOVED**@10.0.0.5:554/cam/realmonitor?channel=3&subtype=1

cameras:
  front_door:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/front_door_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/front_door
        roles:
        - record
    objects:
      track:
      - person
      - cat
      - dog
    motion:
      mask:
      - 187,7,187,25,7,23,7,7
    zones:
      front_porch:
        coordinates: 105,240,352,240,352,93,275,113,224,69,192,22,126,0
      front_lawn:
        coordinates: 219,67,271,104,298,0,185,0

  north_deck:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/north_deck_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/north_deck
        roles:
        - record
    objects:
      track:
      - person
      - cat
      - dog
    motion:
      mask:
      - 331,7,330,23,149,25,149,8

  south_deck:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/south_deck_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/south_deck
        roles:
        - record
    objects:
      track:
      - person
      - cat
      - dog
    motion:
      mask:
      - 333,6,331,24,149,25,151,6
    zones:
      deck:
        coordinates: 0,240,352,240,352,56,323,51,261,32,132,101,0,172
      backyard:
        coordinates: 0,171,89,115,174,72,199,57,226,43,266,28,270,0,232,0,147,0,95,0,93,21,0,59
  front:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/front_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/front
        roles:
        - audio
        - record
    objects:
      track:
      - person
      - bicycle
      - car
      - cat
      - dog
      filters:
        car:
          mask:
          - 352,122,352,0,80,0,54,14,58,38,97,48,156,48
    motion:
      mask:
      - 331,216,333,238,156,239,158,216
    zones:
      front_lawn:
        coordinates: 240,185,277,240,352,240,352,115,267,79,182,50,146,58,107,69,95,101
      front_porch:
        coordinates: 0,134,78,94,54,51,0,73
      driveway_top:
        coordinates: 63,78,149,56,99,28,53,30
      front_patio:
        coordinates: 0,240,254,240,210,172,147,145,108,123,69,99,0,137
    birdseye:
      mode: continuous

  garage:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/garage_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/garage
        roles:
        - audio
        - record
    objects:
      track:
      - person
      - bicycle
      - car
      - cat
      - dog
    motion:
      mask:
      - 180,0,179,21,0,21,0,0
    zones:
      garage_floor:
        coordinates: 40,164,52,202,123,240,352,240,352,98,274,30,104,27
        objects:
        - person
        - bicycle
        - car
        - cat
        - dog
    birdseye:
      mode: continuous

  driveway:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/driveway_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/driveway
        roles:
        - audio
        - record
    objects:
      track:
      - person
      - bicycle
      - car
      - cat
      - dog
      filters:
        car:
          mask:
          - 352,0,352,98,318,74,225,50,141,35,180,240,0,240,0,0
    motion:
      mask:
      - 337,213,332,233,159,233,159,214
    birdseye:
      mode: continuous

  livingroom:
    ffmpeg:
      input_args: preset-rtsp-restream
      inputs:
      - path: rtsp://127.0.0.1:8554/livingroom_sub
        roles:
        - detect
      - path: rtsp://127.0.0.1:8554/livingroom
        roles:
        - record
    objects:
      track:
      - person
      - dog
    motion:
      mask:
      - 341,9,341,23,215,22,214,10
    zones:
      livingroom_area:
        coordinates: 0,240,352,240,352,33,211,38,209,96,49,117,0,174
    birdseye:
      enabled: false

ui:
  live_mode: webrtc


No Coral TPU device detected... even with a new USB cable.. :(

Link to comment

okay got my gpu working in frigate but for some reason the event detection is not working with the tensorrt, I see no events being triggered any ideas why? My config below

 

mqtt:
  enabled: False
detectors:
  tensorrt:
    type: tensorrt
    device: 0
model:
  path: /config/model_cache/tensorrt/yolov7x-640.trt
  width: 640
  height: 640
go2rtc:
  streams:
    bedroom:
      - http://192.168.4.183/flv?port=1935&app=bcs&stream=channel0_main.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD}
cameras:
  bedroom:
    enabled: True
    ffmpeg:
      inputs:
        - path: http://192.168.4.183/flv?port=1935&app=bcs&stream=channel0_main.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD}
          input_args: preset-http-reolink
          roles:
            - record
        - path: http://192.168.4.183/flv?port=1935&app=bcs&stream=channel0_sub.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD}
          input_args: preset-http-reolink
          roles:
            - detect
    detect:
      enabled: True
      width: 640
      height: 480
      fps: 7
    record:
      enabled: True
      retain:
        days: 10
        mode: all
      events:
        retain:
          mode: motion
          default: 20
birdseye:
  enabled: True
  mode: continuous
  restream: False
  width: 640
  height: 480
  quality: 12
ffmpeg:
  hwaccel_args: preset-nvidia-h264
  output_args:
    record: -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -g 300 -c:v hevc_nvenc -preset p7 -tune hq -level 6.2 -tier high -c:a copy

 

and the logs doesn't show any errors:

 

2024-02-10 13:41:15.567309349  [INFO] Preparing Frigate...
2024-02-10 13:41:15.579936723  [INFO] Starting Frigate...
2024-02-10 13:41:17.224436812  [2024-02-10 13:41:17] frigate.app                    INFO    : Starting Frigate (0.13.1-34fb1c2)
2024-02-10 13:41:18.796348475  [2024-02-10 13:41:18] peewee_migrate.logs            INFO    : Starting migrations
2024-02-10 13:41:18.805579722  [2024-02-10 13:41:18] peewee_migrate.logs            INFO    : There is nothing to migrate
2024-02-10 13:41:18.813163234  [2024-02-10 13:41:18] frigate.app                    INFO    : Recording process started: 742
2024-02-10 13:41:18.816235491  [2024-02-10 13:41:18] frigate.app                    INFO    : go2rtc process pid: 99
2024-02-10 13:41:18.841004447  [2024-02-10 13:41:18] frigate.app                    INFO    : Output process started: 753
2024-02-10 13:41:18.862050587  [2024-02-10 13:41:18] detector.tensorrt              INFO    : Starting detection process: 751
2024-02-10 13:41:18.941615303  [2024-02-10 13:41:18] frigate.app                    INFO    : Camera processor started for bedroom: 766
2024-02-10 13:41:18.943448568  [2024-02-10 13:41:18] frigate.app                    INFO    : Capture process started for bedroom: 768
2024-02-10 13:41:20.430493881  [2024-02-10 13:41:20] frigate.detectors.plugins.tensorrt INFO    : Loaded engine size: 317 MiB
2024-02-10 13:41:24.024620445  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +772, GPU +192, now: CPU 1731, GPU 1287 (MiB)
2024-02-10 13:41:24.407048625  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuDNN: CPU +235, GPU +52, now: CPU 1966, GPU 1339 (MiB)
2024-02-10 13:41:24.409465278  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] TensorRT-managed allocation in engine deserialization: CPU +0, GPU +320, now: CPU 0, GPU 320 (MiB)
2024-02-10 13:41:24.422489353  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +8, now: CPU 1649, GPU 1335 (MiB)
2024-02-10 13:41:24.422599790  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuDNN: CPU +0, GPU +8, now: CPU 1649, GPU 1343 (MiB)
2024-02-10 13:41:24.426641410  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] TensorRT-managed allocation in IExecutionContext creation: CPU +0, GPU +222, now: CPU 0, GPU 542 (MiB)
2024-02-10 13:41:24.426747591  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt WARNING : CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars

 

Link to comment
okay got my gpu working in frigate but for some reason the event detection is not working with the tensorrt, I see no events being triggered any ideas why? My config below
 
mqtt: enabled: Falsedetectors: tensorrt:   type: tensorrt   device: 0model: path: /config/model_cache/tensorrt/yolov7x-640.trt width: 640 height: 640go2rtc: streams:   bedroom:     - http://192.168.4.183/flv?port=1935&app=bcs&stream=channel0_main.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD}cameras: bedroom:   enabled: True   ffmpeg:     inputs:       - path: http://192.168.4.183/flv?port=1935&app=bcs&stream=channel0_main.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD}         input_args: preset-http-reolink         roles:           - record       - path: http://192.168.4.183/flv?port=1935&app=bcs&stream=channel0_sub.bcs&user=admin&password={FRIGATE_RTSP_PASSWORD}         input_args: preset-http-reolink         roles:           - detect   detect:     enabled: True     width: 640     height: 480     fps: 7   record:     enabled: True     retain:       days: 10       mode: all     events:       retain:         mode: motion         default: 20birdseye: enabled: True mode: continuous restream: False width: 640 height: 480 quality: 12ffmpeg: hwaccel_args: preset-nvidia-h264 output_args:   record: -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -g 300 -c:v hevc_nvenc -preset p7 -tune hq -level 6.2 -tier high -c:a copy

 
and the logs doesn't show any errors:
 

2024-02-10 13:41:15.567309349  [iNFO] Preparing Frigate...2024-02-10 13:41:15.579936723  [iNFO] Starting Frigate...2024-02-10 13:41:17.224436812  [2024-02-10 13:41:17] frigate.app                    INFO    : Starting Frigate (0.13.1-34fb1c2)2024-02-10 13:41:18.796348475  [2024-02-10 13:41:18] peewee_migrate.logs            INFO    : Starting migrations2024-02-10 13:41:18.805579722  [2024-02-10 13:41:18] peewee_migrate.logs            INFO    : There is nothing to migrate2024-02-10 13:41:18.813163234  [2024-02-10 13:41:18] frigate.app                    INFO    : Recording process started: 7422024-02-10 13:41:18.816235491  [2024-02-10 13:41:18] frigate.app                    INFO    : go2rtc process pid: 992024-02-10 13:41:18.841004447  [2024-02-10 13:41:18] frigate.app                    INFO    : Output process started: 7532024-02-10 13:41:18.862050587  [2024-02-10 13:41:18] detector.tensorrt              INFO    : Starting detection process: 7512024-02-10 13:41:18.941615303  [2024-02-10 13:41:18] frigate.app                    INFO    : Camera processor started for bedroom: 7662024-02-10 13:41:18.943448568  [2024-02-10 13:41:18] frigate.app                    INFO    : Capture process started for bedroom: 7682024-02-10 13:41:20.430493881  [2024-02-10 13:41:20] frigate.detectors.plugins.tensorrt INFO    : Loaded engine size: 317 MiB2024-02-10 13:41:24.024620445  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +772, GPU +192, now: CPU 1731, GPU 1287 (MiB)2024-02-10 13:41:24.407048625  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuDNN: CPU +235, GPU +52, now: CPU 1966, GPU 1339 (MiB)2024-02-10 13:41:24.409465278  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] TensorRT-managed allocation in engine deserialization: CPU +0, GPU +320, now: CPU 0, GPU 320 (MiB)2024-02-10 13:41:24.422489353  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +8, now: CPU 1649, GPU 1335 (MiB)2024-02-10 13:41:24.422599790  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] Init cuDNN: CPU +0, GPU +8, now: CPU 1649, GPU 1343 (MiB)2024-02-10 13:41:24.426641410  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt INFO    : [MemUsageChange] TensorRT-managed allocation in IExecutionContext creation: CPU +0, GPU +222, now: CPU 0, GPU 542 (MiB)2024-02-10 13:41:24.426747591  [2024-02-10 13:41:24] frigate.detectors.plugins.tensorrt WARNING : CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars

 


This was my exact issue on my original post. Somewhere between there and the config I posted yesterday it started working.

Just can’t get my usb coral device to recognize


Sent from my iPhone using Tapatalk
Link to comment
2 hours ago, BrandonG777 said:


This was my exact issue on my original post. Somewhere between there and the config I posted yesterday it started working.

Just can’t get my usb coral device to recognize


Sent from my iPhone using Tapatalk

thinking its my unraid gpu driver version v530.41.03, that I need to update, but can't because I need to update my unraid which is still on 6.12.0-rc3. Feel like unraid will brick itself and the docker containers if I update the OS

Link to comment
3 hours ago, Hellomynameisleo said:

thinking its my unraid gpu driver version v530.41.03, that I need to update, but can't because I need to update my unraid which is still on 6.12.0-rc3. Feel like unraid will brick itself and the docker containers if I update the OS

Yeah I think you're right at the cutoff. I can't remember where I read it but somewhere I think it said v535 was required. 

Link to comment
4 hours ago, BrandonG777 said:

Yeah I think you're right at the cutoff. I can't remember where I read it but somewhere I think it said v535 was required. 

updated to latest version of unraid and nvidia drivers, still doesn't trigger events even though I can see it working on the GPU load for the detector. Might be a frigate bug in this version.

image.png.8108bb29037c5505dbdc0b1de6a8803e.png

 

edit: Got it working now, had to add input_tensor and input_pixel_format

model:

  path: /config/model_cache/tensorrt/yolov7x-640.trt

  input_tensor: nchw

  input_pixel_format: rgb

  width: 640

  height: 640

Edited by Hellomynameisleo
  • Like 2
Link to comment

Well, everything has been going fine for me using GPU detector but I have not been able to get my USB Coral device to detect within the container. Early this morning my entire unRAID hard locked on me. Something I have not seen before plugging in this sketchy USB Coral device in. 


I've mapped /dev/bus/usb and granted the container priviledged access. Beyond this, I can't find anything that I'm missing. Unless someone can point out something I've missed I think I'm throwing in the towel on the USB Coral.

You can see my full config above.

Edited by BrandonG777
pointing out my config above
Link to comment
14 hours ago, BrandonG777 said:

Well, everything has been going fine for me using GPU detector but I have not been able to get my USB Coral device to detect within the container. Early this morning my entire unRAID hard locked on me. Something I have not seen before plugging in this sketchy USB Coral device in. 


I've mapped /dev/bus/usb and granted the container priviledged access. Beyond this, I can't find anything that I'm missing. Unless someone can point out something I've missed I think I'm throwing in the towel on the USB Coral.

You can see my full config above.

Can you test any of the examples using another computer?

https://coral.ai/examples/

 

It is the best way to see if you have a faulty device.

Link to comment

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...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.