[SUPPORT] blakeblackshear - Frigate


Recommended Posts

9 hours ago, digiblur said:

I noticed some issues with the container especially the tmp mapping. It appeared to be causing excessive writes and not set to the right target in the container. I did a PR but haven't heard from the maintainer. Has it been abandoned? 

Sorry I have been busy this weeks, I'll check out your PR today.

 

Edit: merged

Edited by yayitazale
Link to comment
2 hours ago, yayitazale said:

Sorry I have been busy this weeks, I'll check out your PR today.

 

Edit: merged

Thanks! Hopefully others can chime in on the cache target as I noticed nothing was going in the folder until I fixed it. 

 

I also noticed the privileged flag was set to true which I do not recommend unless absolutely necessary. 

Link to comment
5 minutes ago, digiblur said:

Thanks! Hopefully others can chime in on the cache target as I noticed nothing was going in the folder until I fixed it. 

 

I also noticed the privileged flag was set to true which I do not recommend unless absolutely necessary. 

The privileged flag is set on following the instructions of the developer:

 

https://blakeblackshear.github.io/frigate/installation

 

Edit: changed to false as default config.

Edited by yayitazale
Link to comment

Hi,

 

I have a problem not directly related to this docker container I think.

 

I was having issues with some stuff so I thought I should try to reinstall the container, but I can't seem to find it in the "app store" after removing it from my containers. How can I solve this?

Link to comment

Assuming you're talking about frigate, then it's not appearing because of

        "TemplatePath": "/tmp/appFeed/templates/yayitazalesRepository/yayitazale/frigate-amd64.xml",
        "errors": [
            "Opening and ending tag mismatch: Registry line 5 and Container",
            "EndTag: '            "Opening and ending tag mismatch: Container line 2 and Overview",
            "Extra content at the end of the document",
            "Opening and ending tag mismatch: Container line 2 and Overview",
            "Extra content at the end of the document"
        ],

@yayitazale

 

But, it should still appear within Previous Apps

Link to comment
11 hours ago, Squid said:

Assuming you're talking about frigate, then it's not appearing because of



        "TemplatePath": "/tmp/appFeed/templates/yayitazalesRepository/yayitazale/frigate-amd64.xml",
        "errors": [
            "Opening and ending tag mismatch: Registry line 5 and Container",
            "EndTag: '            "Opening and ending tag mismatch: Container line 2 and Overview",
            "Extra content at the end of the document",
            "Opening and ending tag mismatch: Container line 2 and Overview",
            "Extra content at the end of the document"
        ],

@yayitazale

 

But, it should still appear within Previous Apps

pushed an update to correct it

Edited by yayitazale
Link to comment

Hi,

 

I just started to playing with this container and have some question.

 

First is about template.

 

I see this

 

image.thumb.png.dc00f3b1b75fd450fec9d8bf4be927bb.png

 

I think is an error right?

 

Moreover do I need to add intel gpu mapping if I use coral usb? 

 

I can't start container if /dev/dri... is added.

 

 

Edited by Jokerigno
Link to comment
8 hours ago, Jokerigno said:

Hi,

 

I just started to playing with this container and have some question.

 

First is about template.

 

I see this

 

image.thumb.png.dc00f3b1b75fd450fec9d8bf4be927bb.png

 

I think is an error right?

 

Moreover do I need to add intel gpu mapping if I use coral usb? 

 

I can't start container if /dev/dri... is added.

 

 

Yes you are right, this is an error. I deleted the duplicated line, sorry about that.

 

The Intel/AMD GPU is only used for optimizing the performance of the FFMPEG image decoder using integrated GPU for hardware acceleration, reducing your CPU load.

 

The CORAL is used to offloading TensorFlow of the CPU. Is an order of magnitude faster and will reduce your CPU load dramatically.

 

https://blakeblackshear.github.io/frigate/configuration/optimizing

 

If you can't start the container maybe you don't have a hard. acc. compatible CPU, so you just can delete that record of the template and start it without hard. acc. FFMPEG and instead use your CPU to decode image. If you are using low number of cameras you'll not notice much difference.

Edited by yayitazale
Link to comment

First of all this is a great piece of software! I started playing with this yesterday and have my 8 rtsp cameras detecting people, dogs, etc and recording clips.

 

I have 2 questions though:

 

1) I just ordered the Coral M.2 Accelerator A+E key for my Asrock ITX board. I don't use the current wifi card in there so going to replace it with this. How would I go about passing through this M.2 card to this container?

 

2) My other question is should I create a path of /tmp on my host to the container for /tmp/cache? I see there are a bunch of writes being done to this location with temporary ffmpeg created videos. Should they be created on my unraid tmp folder instead? This is how my current Plex container stores it's temporary ffmpeg created videos.

Link to comment
9 hours ago, remati said:

First of all this is a great piece of software! I started playing with this yesterday and have my 8 rtsp cameras detecting people, dogs, etc and recording clips.

 

I have 2 questions though:

 

1) I just ordered the Coral M.2 Accelerator A+E key for my Asrock ITX board. I don't use the current wifi card in there so going to replace it with this. How would I go about passing through this M.2 card to this container?

 

2) My other question is should I create a path of /tmp on my host to the container for /tmp/cache? I see there are a bunch of writes being done to this location with temporary ffmpeg created videos. Should they be created on my unraid tmp folder instead? This is how my current Plex container stores it's temporary ffmpeg created videos.

1) To use  the M.2 coral you need to install the "Coral Accelerator Module Drivers" plugin and then edit this:

 

imagen.thumb.png.b39952c9f00430fb1a9a411ecba2287d.png

to this:

imagen.thumb.png.1880e57d3097955fed858412df5245fc.png

 

 

2) I already changed this on the template this week, look at the hidden settings and you should see this:

 

imagen.thumb.png.b4ec7b592f830e3fa3de51c644a56584.png

 

If you don't see it, delete your container and add it again from the app store.

  • Like 2
Link to comment
On 3/23/2021 at 6:23 AM, mwasserman said:

After seeing the great progress that was being made to get the Mini PCIe Coral working I bought one with an adapter to try my luck. It's not going as smoothly as I had hoped. Maybe someone here can point me in the correct direction or next steps to help debug


For anyone else with this problem, here are two things I tried:

  1. I changed the PCIe slots on the motherboard!
  2. I removed the Coral driver plugin, restarted UNRAID, installed the Coral driver plugin, then restarted UNRAID again.

Before I had my Ableconn PEX-MP117 adapter card in a different slot at first and I was geting the "No EdgeTPU detected" error. Today I tried the above things and I now get "TPU found" instead.

Ableconn card notes:

  • The card has jumpers to change between "PCIe based" and "USB Based" mode. I left mine in the default "PCIe based" position (jumpers on 1-2 & 4-5).
  • I only plugged the Coral mini PCIe board into the Ableconn card. I did not connect anything to the USB port on the back, nor the J5 header on the board.

I'm not sure exactly what worked but, this detectors config is working for me now:

detectors:
  coral:
    type: edgetpu
    device: pci

 

 

  • Like 1
Link to comment

Currently in the process of changing over my system from ZM to Frigate and had some questions I'm not finding answers for.

 

1.

I found an older mention to this in the Github issues as a feature request (#841) but was somewhat confused as it seems to describe older functionality that was added then removed again?

 

Basically I have a camera monitoring my driveway/street looking for cars. Issue is that when motion is detected due to a passing car it almost always picks up a parked car as the highest certainty. Similarly if waving tree branches trigger motion it still assigns the parked car as an event.

 

Is there a way to check for *object* motion and ignore stationary objects? Might be unrealistic to filter out a parked car when a moving car drives behind it but hopefully I can at least get rid of the false positives from wind motion.

 

2. What point of a detected object is associated with the underlying zone? I have a "best image" of a vehicle (stationary car on street when a car passed behind it as above) where the vehicle is entirely within a zone but no zone is listed on the event.

 

3.

Not sure if this is just me or a limitation of Frigate. On the web interface switching "tabs" (ie from cameras to events) doesn't show new events until the webpage is refreshed. Normal?

 

4.

Any convenient way to trigger global object detection for diagnostic purposes over increasing motion sensitivity to an extreme? It would be awesome to just add the bounding boxes, motion, zones, etc to the event video but I assume that's not an option yet.

 

Thanks,

-Nick

Link to comment
14 hours ago, BlackAndChrome said:

Currently in the process of changing over my system from ZM to Frigate and had some questions I'm not finding answers for.

 

1.

I found an older mention to this in the Github issues as a feature request (#841) but was somewhat confused as it seems to describe older functionality that was added then removed again?

 

Basically I have a camera monitoring my driveway/street looking for cars. Issue is that when motion is detected due to a passing car it almost always picks up a parked car as the highest certainty. Similarly if waving tree branches trigger motion it still assigns the parked car as an event.

 

Is there a way to check for *object* motion and ignore stationary objects? Might be unrealistic to filter out a parked car when a moving car drives behind it but hopefully I can at least get rid of the false positives from wind motion.

 

2. What point of a detected object is associated with the underlying zone? I have a "best image" of a vehicle (stationary car on street when a car passed behind it as above) where the vehicle is entirely within a zone but no zone is listed on the event.

 

3.

Not sure if this is just me or a limitation of Frigate. On the web interface switching "tabs" (ie from cameras to events) doesn't show new events until the webpage is refreshed. Normal?

 

4.

Any convenient way to trigger global object detection for diagnostic purposes over increasing motion sensitivity to an extreme? It would be awesome to just add the bounding boxes, motion, zones, etc to the event video but I assume that's not an option yet.

 

Thanks,

-Nick

 

As far as I'm not the developer of the APP, I'm not sure about any of the questions you have, but I'll try to help. Anyway, you can ask these questions on the Github page.

 

1. I think currently it doesn't ignore stationary objects because in the end you are passing static frames to TFlite and this can't really know witch of the 2 cars is the one tha't it's moving. To avoid false positives from wind motion you have to adjust the values of the detection as you can read here: https://blakeblackshear.github.io/frigate/configuration/false_positives

 

On next releases maybe this funtionality will be added: https://github.com/blakeblackshear/frigate/issues/375

 

2. I think zones are not linked to the detected objects for now.

 

3. Not sure about that

 

4. There is no way to activate a global detectión, you must specify the objetcs types you want to detec one by one. You can see the boxes on the video stream for debbuging purposes, you need to navigate to http://:5000/api/<camera_name>?bbox=1

 

https://blakeblackshear.github.io/frigate/usage/api/#apicamera_name

 

 

Link to comment

Hi All,

 

Considering moving from BI+Deepstack to Frigate so I installed that docker

First of all, thanks a lot for the work because I had it up and running in less than 10min. Installation was really ez.

 

Now I am running some few tests (CPU mode) and I have to say the detection is pretty good.

 

However, the whole thing will work for 10-15min and then I noticed some issues.

My camera stream becomes messy (see attached) and I have the errors below in the log (it's just an extract)

 

 

[h264 @ 0x562cc0734d00] left block unavailable for requested intra mode
[h264 @ 0x562cc0734d00] error while decoding MB 0 4, bytestream 954579
[h264 @ 0x562cc0734d00] left block unavailable for requested intra mode
[h264 @ 0x562cc0734d00] error while decoding MB 0 114, bytestream 16535
[h264 @ 0x55fa44f00d00] error while decoding MB 158 17, bytestream -31
[h264 @ 0x55fa44f00d00] cabac decode of qscale diff failed at 109 2
[h264 @ 0x55fa44f00d00] error while decoding MB 109 2, bytestream 1134689
[h264 @ 0x55fa44f00d00] error while decoding MB 104 5, bytestream -9
[h264 @ 0x559eba2b6d00] error while decoding MB 115 13, bytestream -20
[h264 @ 0x559eba2b6d00] cabac decode of qscale diff failed at 19 66
[h264 @ 0x559eba2b6d00] error while decoding MB 19 66, bytestream 27113
[h264 @ 0x55dc86788d00] error while decoding MB 144 2, bytestream -31
[h264 @ 0x55dc86788d00] cabac decode of qscale diff failed at 55 50
[h264 @ 0x55dc86788d00] error while decoding MB 55 50, bytestream 40441
[h264 @ 0x556c09134d00] error while decoding MB 54 17, bytestream -21

 

The stream comes from a reolink camera and when I have a look on it in BlueIris, it is crystal clear

I believe this is an FFMPEG issuer but I have no idea how to fix it

 

 

 

Any idea what could be the issue?

canvas.png

Edited by mathgoy
Link to comment
4 hours ago, mathgoy said:

Hi All,

 

Considering moving from BI+Deepstack to Frigate so I installed that docker

First of all, thanks a lot for the work because I had it up and running in less than 10min. Installation was really ez.

 

Now I am running some few tests (CPU mode) and I have to say the detection is pretty good.

 

However, the whole thing will work for 10-15min and then I noticed some issues.

My camera stream becomes messy (see attached) and I have the errors below in the log (it's just an extract)

 

 





[h264 @ 0x562cc0734d00] left block unavailable for requested intra mode
[h264 @ 0x562cc0734d00] error while decoding MB 0 4, bytestream 954579
[h264 @ 0x562cc0734d00] left block unavailable for requested intra mode
[h264 @ 0x562cc0734d00] error while decoding MB 0 114, bytestream 16535
[h264 @ 0x55fa44f00d00] error while decoding MB 158 17, bytestream -31
[h264 @ 0x55fa44f00d00] cabac decode of qscale diff failed at 109 2
[h264 @ 0x55fa44f00d00] error while decoding MB 109 2, bytestream 1134689
[h264 @ 0x55fa44f00d00] error while decoding MB 104 5, bytestream -9
[h264 @ 0x559eba2b6d00] error while decoding MB 115 13, bytestream -20
[h264 @ 0x559eba2b6d00] cabac decode of qscale diff failed at 19 66
[h264 @ 0x559eba2b6d00] error while decoding MB 19 66, bytestream 27113
[h264 @ 0x55dc86788d00] error while decoding MB 144 2, bytestream -31
[h264 @ 0x55dc86788d00] cabac decode of qscale diff failed at 55 50
[h264 @ 0x55dc86788d00] error while decoding MB 55 50, bytestream 40441
[h264 @ 0x556c09134d00] error while decoding MB 54 17, bytestream -21

 

The stream comes from a reolink camera and when I have a look on it in BlueIris, it is crystal clear

I believe this is an FFMPEG issuer but I have no idea how to fix it

 

 

 

Any idea what could be the issue?

canvas.png

 

After some further investigations, I found out that it's related to the Reolink Cameras I use.

 

It's recommended to use a RTMP stream instead of an RTSP.

 

For those who may have the same issue, here is the link format for RTMP streams on reolink cameras:

rtmp://IPADRESS/bcs/channel0_main.bcs?channel=0&stream=0&user=USER&password=PASSWORD

 

 

You will find below the block that worked for my Reolink camera (with the input_args)

cameras:
  descente:
    ffmpeg:      
      inputs:
 
        - path: rtmp://IP_ADDRESS/bcs/channel0_main.bcs?channel=0&stream=0&user=YOURUSER&password=YOURPASSWORD
          roles:
            - detect
            - clips
            - record
      input_args:
        - -avoid_negative_ts
        - make_zero
        - -fflags
        - nobuffer
        - -flags
        - low_delay
        - -strict
        - experimental
        - -fflags
        - +genpts+discardcorrupt
        - -rw_timeout
        - '5000000'
        - -use_wallclock_as_timestamps
        - '1'    
    width: 2560
    height: 1920
    fps: 6

 

Edited by mathgoy
Link to comment

HI. thanks for making such a great software! 

I tried to play around with label mapping by changing "person" to "human". i have downloaded the coco_label.txt and changed its name to labelmap.txt and placed it in "/mnt/user/appdata/frigate/labelmap/".

In this process I had troubles with making it work. I have tried the two following ways of mounting the labelmap.txt file without any luck:

image.png.710720e40173b8833997d0cea3e919ab.png

With the /labelmap.txt unraid gives the error "bad parameter"

image.png.c78ea2688a15f974affdacba59a332e1.png

 

In my config.yaml file i have added the "objects" variable like so:

image.png.2f93fb248ce1f3434cc4cc5b032b726e.png

Link to comment
20 minutes ago, kasperhonore said:

HI. thanks for making such a great software! 

I tried to play around with label mapping by changing "person" to "human". i have downloaded the coco_label.txt and changed its name to labelmap.txt and placed it in "/mnt/user/appdata/frigate/labelmap/".

In this process I had troubles with making it work. I have tried the two following ways of mounting the labelmap.txt file without any luck:

image.png.710720e40173b8833997d0cea3e919ab.png

With the /labelmap.txt unraid gives the error "bad parameter"

image.png.c78ea2688a15f974affdacba59a332e1.png

 

In my config.yaml file i have added the "objects" variable like so:

image.png.2f93fb248ce1f3434cc4cc5b032b726e.png

 

The first try is the way but you are trying to match /labelmap.txt with a folder labelmap.

 

The correct way it should be to match a file with a file:

 

container path : /labelmap.txt

host path: /mnt/user/appdata/frigate/labelmap/labelmap.txt

 

You can see it on the docs: https://blakeblackshear.github.io/frigate/configuration/objects#customizing-the-labelmap

 

Cheers

 

Edited by yayitazale
Link to comment
3 minutes ago, yayitazale said:

 

The first try is the way but you are trying to match /labelmap.txt with a folder labelmap.

 

The correct way it should be to match a file with a file:

 

container path : /labelmap.txt

host path: /mnt/user/appdata/frigate/labelmap/labelmap.txt

 

You can see it on the docs: https://blakeblackshear.github.io/frigate/configuration/objects#customizing-the-labelmap

 

Cheers

 

Thanks for the help! It works now

  • Like 1
Link to comment

 

Hi all, hope someone can please help

 

i have unraid and this docker working perfectly on one system but its a real power hog so I have been trying to get this working on my hp proliant nl36 which only has a AMD Athlon II, 2 core 1.30GHz CPU.

 

From what I have found out this CPU does not support AVX, and I think that is what is causing this error below,

 

Iv also asked on the Home Assistant forum post for Frigate but I did not get any replies.

 

Any suggestions appreciated, thanks

 

https://github.com/blakeblackshear/frigate/issues/49

 

```

* Starting nginx nginx
...done.
Fatal Python error: Illegal instruction

Current thread 0x00001553e214a740 (most recent call first):
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 1101 in create_module
File "<frozen importlib._bootstrap>", line 556 in module_from_spec
File "<frozen importlib._bootstrap>", line 657 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap>", line 1042 in _handle_fromlist
File "/usr/local/lib/python3.8/dist-packages/tflite_runtime/interpreter.py", line 36 in <module>
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 783 in exec_module
File "<frozen importlib._bootstrap>", line 671 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "/opt/frigate/frigate/edgetpu.py", line 15 in <module>
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 783 in exec_module
File "<frozen importlib._bootstrap>", line 671 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "/opt/frigate/frigate/app.py", line 19 in <module>
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 783 in exec_module
File "<frozen importlib._bootstrap>", line 671 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "/opt/frigate/frigate/__main__.py", line 7 in <module>
File "/usr/lib/python3.8/runpy.py", line 87 in _run_code
File "/usr/lib/python3.8/runpy.py", line 194 in _run_module_as_main

```

image.thumb.png.5d2d9402f9a9d973af612dbde908cdc0.png

Edited by Allanon
Link to comment
9 minutes ago, Allanon said:

 

Hi all, hope someone can please help

 

i have unraid and this docker working perfectly on one system but its a real power hog so I have been trying to get this working on my hp proliant nl36 which only has a AMD Athlon II, 2 core 1.30GHz CPU.

 

From what I have found out this CPU does not support AVX, and I think that is what is causing this error below,

 

Iv also asked on the Home Assistant forum post for Frigate but I did not get any replies.

 

Any suggestions appreciated, thanks

 

```

* Starting nginx nginx
...done.
Fatal Python error: Illegal instruction

Current thread 0x00001553e214a740 (most recent call first):
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 1101 in create_module
File "<frozen importlib._bootstrap>", line 556 in module_from_spec
File "<frozen importlib._bootstrap>", line 657 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap>", line 1042 in _handle_fromlist
File "/usr/local/lib/python3.8/dist-packages/tflite_runtime/interpreter.py", line 36 in <module>
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 783 in exec_module
File "<frozen importlib._bootstrap>", line 671 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "/opt/frigate/frigate/edgetpu.py", line 15 in <module>
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 783 in exec_module
File "<frozen importlib._bootstrap>", line 671 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "/opt/frigate/frigate/app.py", line 19 in <module>
File "<frozen importlib._bootstrap>", line 219 in _call_with_frames_removed
File "<frozen importlib._bootstrap_external>", line 783 in exec_module
File "<frozen importlib._bootstrap>", line 671 in _load_unlocked
File "<frozen importlib._bootstrap>", line 975 in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 991 in _find_and_load
File "/opt/frigate/frigate/__main__.py", line 7 in <module>
File "/usr/lib/python3.8/runpy.py", line 87 in _run_code
File "/usr/lib/python3.8/runpy.py", line 194 in _run_module_as_main

```

image.thumb.png.5d2d9402f9a9d973af612dbde908cdc0.png

 

Your CPU is not supported, the only way to make it work is using the modified version of the hass addon:

 

https://github.com/blakeblackshear/frigate/issues/695

https://github.com/pdecat/frigate-hass-addons

  • Like 1
Link to comment
2 hours ago, yayitazale said:

 

Your CPU is not supported, the only way to make it work is using the modified version of the hass addon:

 

https://github.com/blakeblackshear/frigate/issues/695

https://github.com/pdecat/frigate-hass-addons

Hi thanks for that but when i go to https://github.com/pdecat/frigate-hass-addons

 

I'm getting a 404 page not found    https://github.com/pdecat/frigate-hass-addons/blob/main/frigate_beta_old/README.md

 

Sorry if I am doing something daft 😞

image.png

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.