Codeproject Ai

[scuzzie2k]

Since the update to 2.1.6 my Blueiris was only detecting Cars, nothing else, so i removed and re-installed it, I get errors all over my CodeProject AI when I try to enable GPU on Object Detection (YOLOv5 6.2)

21:12:46:face.py: Traceback (most recent call last):
21:12:46:face.py: File "C:\Program Files\CodeProject\AI\modules\FaceProcessing\intelligencelayer\face.py", line 34, in
21:12:46:face.py: import torch
21:12:46:face.py: ModuleNotFoundError: No module named 'torch'

I have removed it all again, deleted the directory in program files and programdata.

 

[Kujo]

Since the update to 2.1.6 my Blueiris was only detecting Cars, nothing else, so i removed and re-installed it, I get errors all over my CodeProject AI when I try to enable GPU on Object Detection (YOLOv5 6.2)

21:12:46:face.py: Traceback (most recent call last):
21:12:46:face.py: File "C:\Program Files\CodeProject\AI\modules\FaceProcessing\intelligencelayer\face.py", line 34, in
21:12:46:face.py: import torch
21:12:46:face.py: ModuleNotFoundError: No module named 'torch'

I have removed it all again, deleted the directory in program files and programdata.

Same here did you get this resolved

 

[smiticans]

Same here did you get this resolved

Follow the last post in this thread. Be Careful with CodeProject AI 2.1.0

Make sure to shutdown BI, uninstall codeproject, delete the folders specified in the above thread, reboot the computer then reinstall codeproject. Once CPAI says its's done installing, open the dashboard and keep an eye on the logs. It takes another 5 to 10 minutes to install all of the modules. Once it's done installing the modules, then you can run Blue Iris again. I had this same issue and BI would cause the install to become corrupt because I would run BI before CPAI was done installing.

 

[johndball]

Just a heads up in case somebody finds this thread: I ran into the same issue but blowing out the install and starting from scratch didn't resolve it.
Solution: I had two issues going on: 1) I have SSL/TLS MITM inspection at my edge/gateway. I needed to exclude *.pytorch.org from inspection as the module was throwing an error on download. 2) I installed Torch manually (as a test) to rule out any other issues. I used

pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

from the

C:\Program Files\CodeProject\AI\modules\ObjectDetectionYOLOv8\bin\windows\python39\venv\Scripts

directory and it went without issue.

Face Processing (1.12.3) starts with GPU (CUDA) and doesn't crash.