5.5.8 - June 13, 2022 - Code Project’s SenseAI Version 1 - See V2 here https://ipcamtalk.com/threads/codeproject-ai-version-2-0.68030/

Ok I'm having problems.

A few days ago, a lot of my alerts started being cancelled and returning 'AI not responding'.
I rebooted the system and all was working again. No biggie, I thought.

But now it's doing it again.

1667122644515.png

It doesn't happen to all of them, and it seems pretty arbitrary, but it's consistently infuriating.

When I check task manager, I realize that at least 38 instances of Python are running (!).
1667124738270.png

How is this possible?
I'm running 1.6.7-Beta.

Any ideas?
 
Last edited:
Ok I'm having problems.

A few days ago, a lot of my alerts started being cancelled and returning 'AI not responding'.
I rebooted the system and all was working again. No biggie, I thought.

But now it's doing it again.

View attachment 144247

It doesn't happen to all of them, and it seems pretty arbitrary, but it's consistently infuriating.

When I check task manager, I realize that at least 38 instances of Python are running (!).
View attachment 144249

How is this possible?
I'm running 1.6.7-Beta.

Any ideas?
Post a screenshot of your BI main AI settings also CP.AI dashboard
 
Does anyone's setup with Blue Iris constantly detect objects? I'm seeing detection every 2-3 seconds and when there is an actual object to detect, BI notifications are showing 4+objects detected over one object (For example, 1 person, 4 detections).
 
Last edited:
Since you are using a CPU for CP.AI you can try setting the analyze one each to 1sec so it gives CP.AI time to process the image.

I don't understand why it's timing out though? Normally it returns ~100ms analysis times, and these errors also happen at night, when there's barely any activity, beyond the rare car that drives past.
 
I don't understand why it's timing out though? Normally it returns ~100ms analysis times, and these errors also happen at night, when there's barely any activity, beyond the rare car that drives past.
Check your BI log, are you seeing timeouts along with a restart.
 
I finally upgraded from DeepStack today to CodeProject.AI and its seems to be working fine.

However I note that the CodeProject.AI.Server.exe is writing a lot of temp files which appear to be the ASP.NET core request files it writes for requests over 30KB unless buffering is enabled. For example below are some of the writes to one temp file for a single image analysis call below:

(Note: I had already redirected them to another drive using the 'ASPNETCORE_TEMP ' env variable to test before taking this screenshot, they're usually in the Window temp folder):

Clipboard Image.jpg

Each temp file is deleted at the end of the analysis request so you won't see them usually, but it isn't an optimum way to work if it can be avoided.

I'll setup the CodeProject AI source code in dev mode when I have time to investigate if its as simple as enabling the ASP.NET request memory buffering feature with a better limit, in the mean time any thoughts @MikeLud1 ?
 
When I first installed AI months ago it worked fine. If I received a push notification it would most likely be a car, person. But for the last 4 weeks I get a push notification even when AI say’s ”nothin found”. How can I get only notifications on only car, person, etc. thanks!
 
I am currently running DS on my BI server using your custom models Mike.

I am trying to move to running CodeProject in a Docker container on my Unraid server, which is of course, a separate machine. I got the Docker container up and running, and I have successfully got my BI server pointed to the Docker and it is working. However, I can't figure out how to get custom models working when CodeProject is running on a separate machine in a Docker. Does anyone have any ideas or can they help me with this? Really want to get the dark and combined custom models up and running as detection is much better.

edit The biggest issue is I can't figure out where or how to install the custom models.
 
Last edited:
I am currently running DS on my BI server using your custom models Mike.


edit The biggest issue is I can't figure out where or how to install the custom models.
FWIIW: For testing, I am using AI via Docker in a Linux VM. I created a Samba share of the folder with the models and put that in the AI on BI:
\\192.168.50.17\cuda
1667587286703.png
 
  • Like
Reactions: MikeLud1
Hmmm. Ok. I'll have to play with that. My biggest issue is the docker image didn't install the app info in the usual directory on my Unraid. So I can't figure out what folder to put the models in so that CodeProject will recognize them. Once I get that part then sharing the folder and accessing it in BI is simple.