CodeProject.AI Version 2.0

I am seeing a strange anomaly with v2.19. Initially .NET does not load the GPU which in my case is an RTX 1060. When I forced GPU mode, it switched momentarily to ML and then switched back to CPU.

I left it in CPU mode and after what seemed a minute or so it finally switched to ML.

My take presently is that 2.18 is more stable. Needless to say your mileage may vary.
 
  • Like
Reactions: David L
Hell I'm still running 2.08, seems every version afterwards has just caused problems or broken things and hasn't improved anything or added any new features. Not upgrading past 2.08 until there's a substantial reason to.
My experience is that 2.18 is much more efficient than all earlier releases. My GPU load for 8 cameras dropped substantially when I upgraded from 2.17 to 2.18. I'm presently considering reverting to 2.18 from 2.19
 
Last edited:
  • Like
Reactions: David L
Anyone else expericning issues with the latest CPAI 2.1.9 on docker? Nothing is being confirmed in the alerts. I noticed in the clip thumbnail it's saying "AI: Timeout".

1684547304310.png


1684547025281.png

1684547043610.png


1684629489615.png
1684629393544.png




1684628728235.png

1684628764926.png
 
Last edited:
Anyone else expericning issues with the latest CPAI 2.1.9 on docker? Nothing is being confirmed in the alerts. I noticed in the clip thumbnail it's saying "AI: Timeout".

View attachment 163557


View attachment 163555

View attachment 163556





View attachment 163649

View attachment 163650
The issue is you are using two Object Detection modules, you should only use one Object Detection module. Try disabling one of them, since you have only a CPU try the Object Detection (YOLOv5 .NET)

1684629341131.png
 
The issue is you are using two Object Detection modules, you should only use one Object Detection module. Try disabling one of them, since you have only a CPU try the Object Detection (YOLOv5 .NET)

View attachment 163651

I did have just the .NET and LPR running before. Not sure how the 6.2 got restarted. After stopping the YOLO 6.2 I am still having "AI:timeouts" in the clips.

1684694463619.png

1684694509854.png


1684694533532.png
 
Yeah mine looks like that too:

1684713588593.png1684713617004.png
 
I just reverted back to 2.0.8 in docker and everything is working again. No time out issues.

1684870447152.png

1684870573118.png
 
  • Like
Reactions: David L
Anyone running CPAI 2.1.9 and BI 5.7.7.2 seeing a runtime error in CPAI? So far it doesn't appear to be impacting AI detection or speed. I see the error at CPAI start but not again. I've tried uninstalling and reinstalling CPAI, same error.

Full log:

08:04:05:Object Detection (YOLOv5 6.2): [RuntimeError] : Traceback (most recent call last):
File "C:\Program Files\CodeProject\AI\modules\ObjectDetectionYolo\detect.py", line 162, in do_detection
det = detector(img, size=640)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\autograd\grad_mode.py", line 28, in decorate_context
return func(*args, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\yolov5\models\common.py", line 705, in forward
y = self.model(x, augment=augment) # forward
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\yolov5\models\common.py", line 515, in forward
y = self.model(im, augment=augment, visualize=visualize) if augment or visualize else self.model(im)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\Lib\site-packages\yolov5\models\yolo.py", line 209, in forward
return self._forward_once(x, profile, visualize) # single-scale inference, train
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\Lib\site-packages\yolov5\models\yolo.py", line 121, in _forward_once
x = m(x) # run
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\Lib\site-packages\yolov5\models\yolo.py", line 74, in forward
xy = (xy * 2 + self.grid) * self.stride # xy
RuntimeError: The size of tensor a (60) must match the size of tensor b (48) at non-singleton dimension 2
 
Anyone running CPAI 2.1.9 and BI 5.7.7.2 seeing a runtime error in CPAI? So far it doesn't appear to be impacting AI detection or speed. I see the error at CPAI start but not again. I've tried uninstalling and reinstalling CPAI, same error.

Full log:

08:04:05:Object Detection (YOLOv5 6.2): [RuntimeError] : Traceback (most recent call last):
File "C:\Program Files\CodeProject\AI\modules\ObjectDetectionYolo\detect.py", line 162, in do_detection
det = detector(img, size=640)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\autograd\grad_mode.py", line 28, in decorate_context
return func(*args, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\yolov5\models\common.py", line 705, in forward
y = self.model(x, augment=augment) # forward
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\yolov5\models\common.py", line 515, in forward
y = self.model(im, augment=augment, visualize=visualize) if augment or visualize else self.model(im)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\Lib\site-packages\yolov5\models\yolo.py", line 209, in forward
return self._forward_once(x, profile, visualize) # single-scale inference, train
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\Lib\site-packages\yolov5\models\yolo.py", line 121, in _forward_once
x = m(x) # run
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "C:\Program Files\CodeProject\AI\runtimes\bin\windows\python37\venv\Lib\site-packages\yolov5\models\yolo.py", line 74, in forward
xy = (xy * 2 + self.grid) * self.stride # xy
RuntimeError: The size of tensor a (60) must match the size of tensor b (48) at non-singleton dimension 2
What model are you using
 
What model are you using

YOLOv5 6.2, large. The .NET module is stopped. Using "ipcam-combined" for my day profiles, and "ipcam-combined,ipcam-dark" for night, and "license-plate" for LPR.

I've tried rolling back to CPAI 2.1.8 and BI 5.7.6.8, still the same error. It appears the error is a result of BI starting, w/o BI, CPAI starts w/o the error.
 
Last edited:
Hey guys, so I installed cuda driver 516.94, cuda toolkit 11.7, manually installed cuDNN 8.5.0.96 sense I tried codeprojects cuDNN script it would just pull a cmd window up for a split second have no writing in it then just shut "dont know if thats normal or if im even using that tool correctly". Then I installed codeprojects ai 2.1.9, enabled GPU in the ai tab of blue iris and my codeprojects does say its using GPU (CUDA) but when i look at task manager and watch the processes of Python its using mostly CPU power and very little GPU power (using a 1660ti). Now im just wondering if thats normal or if i installed something wrong because back when i had DeepStack GPU Version running i could watch python and it would use only GPU power if im not mistaken. Also what is (Half Precision) sorry if this isn't the right place for this question but been searching the internet for the past couple of days but cant find anything if this is an issue or if this is just normal for codeprojects ai.

Thanks in advance!!!
 

Attachments

  • Codeproject AI Usage 1.PNG
    Codeproject AI Usage 1.PNG
    165.2 KB · Views: 26
  • Codeproject AI Usage.PNG
    Codeproject AI Usage.PNG
    172.5 KB · Views: 29
  • Codeproject AI Server.PNG
    Codeproject AI Server.PNG
    143.9 KB · Views: 34
  • Blue Iris.PNG
    Blue Iris.PNG
    89.4 KB · Views: 33