CodeProject.AI Version 2.0

I'm using v2.4.5 with the google coral tpu however I don't get AI objects on the tuning and testing in BlueIris does anyone know why? Normal detections and alerts work fine.
 
I have a media server running Plex sitting right next to my BI server so I set it up with CPAI 2.4.5 running mesh. I only have four cameras running but curious how this will work.Screenshot 2023-12-17 064350.jpg
 
  • Like
Reactions: David L
BORIStheBLADE said:
I have a media server running Plex sitting right next to my BI server so I set it up with CPAI 2.4.5 running mesh. I only have four cameras running but curious how this will work.View attachment 180308
Click to expand...

Ok well after setting this up and not touching anything I took a walk and set off my cameras but the second PC set up with mesh didn't process anything. What I did notice is with setting both to the "large" model I'm getting faster processes than I did with the single PC running the "small " model. I' don't understand how thats happened. I have to clean my house before I touch this again. Got to keep the wife happy.
 
  • Haha
Reactions: fenderman
105437 said:
That worked! It also fixed the issue where I couldn't install YOLOv8 to test with.

Also, where can I find YOLO test images to add to the TestData folder for CPAI to run benchmarks and vision detections? Currently my TestData folder has these two images.

View attachment 180309
Click to expand...
Attached are all the TestData images
 

Attachments

  • Like
Reactions: 105437
MikeLud1 said:
Attached are all the TestData images
Click to expand...
Using the pexels-maria-orlova-4947381 image on my PC here are the measurements for ipcam-general.

# Label Confidence
0 person 81%
1 car 80%
2 traffic light 63%
3 truck 60%
4 traffic light 59%
5 car 57%
6 truck 54%
7 traffic light 52%
8 truck 52%
9 car 50%
10 truck 48%
11 car 47%
12 traffic light 41%
Processed by ObjectDetectionYOLOv5-6.2
Processed on localhost
Analysis round trip 153 ms
Processing 140 ms
Inference 135 ms


# Label Confidence
0 car 89%
1 person 83%
2 traffic light 59%
3 truck 59%
4 truck 54%
5 car 53%
6 car 46%
7 truck 44%
8 traffic light 44%
9 car 42%
10 truck 41%
Processed by ObjectDetectionYOLOv8
Processed on localhost
Analysis round trip 1218 ms
Processing 1204 ms
Inference 1070 ms
 
105437 said:
Using the pexels-maria-orlova-4947381 image on my PC here are the measurements for ipcam-general.

# Label Confidence
0 person 81%
1 car 80%
2 traffic light 63%
3 truck 60%
4 traffic light 59%
5 car 57%
6 truck 54%
7 traffic light 52%
8 truck 52%
9 car 50%
10 truck 48%
11 car 47%
12 traffic light 41%
Processed by ObjectDetectionYOLOv5-6.2
Processed on localhost
Analysis round trip 153 ms
Processing 140 ms
Inference 135 ms


# Label Confidence
0 car 89%
1 person 83%
2 traffic light 59%
3 truck 59%
4 truck 54%
5 car 53%
6 car 46%
7 truck 44%
8 traffic light 44%
9 car 42%
10 truck 41%
Processed by ObjectDetectionYOLOv8
Processed on localhost
Analysis round trip 1218 ms
Processing 1204 ms
Inference 1070 ms
Click to expand...

Where is this report located in the system? Thx
 
105437 said:
Using the pexels-maria-orlova-4947381 image on my PC here are the measurements for ipcam-general.

# Label Confidence
0 person 81%
1 car 80%
2 traffic light 63%
3 truck 60%
4 traffic light 59%
5 car 57%
6 truck 54%
7 traffic light 52%
8 truck 52%
9 car 50%
10 truck 48%
11 car 47%
12 traffic light 41%
Processed by ObjectDetectionYOLOv5-6.2
Processed on localhost
Analysis round trip 153 ms
Processing 140 ms
Inference 135 ms


# Label Confidence
0 car 89%
1 person 83%
2 traffic light 59%
3 truck 59%
4 truck 54%
5 car 53%
6 car 46%
7 truck 44%
8 traffic light 44%
9 car 42%
10 truck 41%
Processed by ObjectDetectionYOLOv8
Processed on localhost
Analysis round trip 1218 ms
Processing 1204 ms
Inference 1070 ms
Click to expand...
It looks like you tested with default model, ipcam-general should look like the below

# Label Confidence
0 vehicle 87%
1 person 83%
2 vehicle 81%
3 vehicle 80%
4 vehicle 80%
5 vehicle 55%
6 person 52%
7 vehicle 48%
Processed by ObjectDetectionYOLOv5Net
Processed on PC-NVR
Analysis round trip 23 ms
Processing 19 ms
Inference 19 ms
 
  • Like
Reactions: actran
MikeLud1 said:
Let me sleep on it.
Click to expand...

Last night I uninstalled CUDA 11.8 and installed CUDA 12.x. uninstalled the modules, then reinstalled the modules not using cache then rebooted. Had to modify the module.json file again to lower the compute minimum for my GPU. This brought a new issue, where the alpr.py timed out as it couldnt find a file. I dont remember the exact file. I gave up and reverted everything back to CUDA 11.8.

Any thoughts on what to do next?
 
prsmith777 said:
Last night I uninstalled CUDA 11.8 and installed CUDA 12.x. uninstalled the modules, then reinstalled the modules not using cache then rebooted. Had to modify the module.json file again to lower the compute minimum for my GPU. This brought a new issue, where the alpr.py timed out as it couldnt find a file. I dont remember the exact file. I gave up and reverted everything back to CUDA 11.8.

Any thoughts on what to do next?
Click to expand...
On previous versions did you have CUDA working for the ALPR module? I am try figure out if the current ALPR version might be the issue.
 
You must log in or register to reply here.
Top Bottom