Steve_Texas
n3wb
Update on this: This also no longer works on the R710's CPU. I get the same behavior/error when trying to run ALPR on the R710 as I do on my i7 machine. This absolutely worked in previous versions of CPAI and ALPR.
Blue Iris and CodeProject.AI ALPR
- Thread starter MikeLud1
- Start date
Update on this: This also no longer works on the R710's CPU. I get the same behavior/error when trying to run ALPR on the R710 as I do on my i7 machine. This absolutely worked in previous versions of CPAI and ALPR.
Hmm, Interesting - the Nehalem CPU gen of both desktop + servers appears to be impacted.
Ya. One test I tried was CP 2.1.8 , forget which ALPR version it defaulted to install. But it did start ALPR CPU and didn't crash. But no CUDA there either.
Perhaps its due to newer versions of Paddle bundled with ALPR module. Perhaps Mr. Lud knows the history of paddle versions against ALPR module versions? Maybe earlier paddle version could be used along side current ALPR version? I'm out of my element...
Ya. One test I tried was CP 2.1.8 , forget which ALPR version it defaulted to install. But it did start ALPR CPU and didn't crash. But no CUDA there either.
Perhaps its due to newer versions of Paddle bundled with ALPR module. Perhaps Mr. Lud knows the history of paddle versions against ALPR module versions? Maybe earlier paddle version could be used along side current ALPR version? I'm out of my element...
After updating to 2.5.6 I stopped getting plate recognition. In the alert list I think it was showing something about a 500 error. Anyone else have similar issues? Also, what are the latest recommended config details in the AI config for the LPR. I’ve attached my details and also a sample photo. I’ve since reverted back to 2.5.4 but was curious if there was some config modification needed after upgrading. Also, most of the time the motion and LPR works great but sometimes it seems like it will miss just a single vehicle completely. Like it’s asleep at the wheel. No canceled alert, it wasn’t already triggered. Daytime. Easy catch and it just completely misses it. Any suggestions?
Attachments
I managed to update CP.AI (2.6.2) and ALPR (3.1.0) and now I do have one camera capturing cars going up and down the street.
The behavior didn't change.
If I go into CP.AI Project Explorer and use "Read License Plates" on an image where ALPR can't recognize anything, I get the expected output: "No predictions returned".
But if I choose any image which does have a license plate, the module shuts down.
This is what Event Viewer shows:
Windows 11, CP.AI 2.6.2, ALPR 3.1.0, latest Nvidia Drivers installed on RTX 3090 GPU, the works.
Using YOLOv8 for Object Detection.
ALPR data:
I don't know what else to do to troubleshoot it. Needless to say, the module never worked, even when checking known static images with clearly visible license plates.
The behavior didn't change.
If I go into CP.AI Project Explorer and use "Read License Plates" on an image where ALPR can't recognize anything, I get the expected output: "No predictions returned".
But if I choose any image which does have a license plate, the module shuts down.
Code:
19:53:29:Object Detection (YOLOv8): Detecting using license-plate
19:53:29:Response rec'd from Object Detection (YOLOv8) command 'custom' (...fcd391) ['Found Plate'] took 80ms
19:53:33:Module ALPR has shutdown
19:53:33:ALPR_adapter.py: has exitedThis is what Event Viewer shows:
Code:
Faulting application name: python.exe, version: 3.9.6150.1013, time stamp: 0x60d9eb23
Faulting module name: cudnn_ops_infer64_8.dll, version: 6.14.11.11070, time stamp: 0x62e9c2eb
Exception code: 0xc0000409
Fault offset: 0x00000000003298d9
Faulting process id: 0x0xC7C
Faulting application start time: 0x0x1DA88420F6018A7
Faulting application path: C:\Program Files\CodeProject\AI247\runtimes\bin\windows\python39\python.exe
Faulting module path: C:\Program Files\NVIDIA\CUDNN\v8.5\bin\cudnn_ops_infer64_8.dll
Report Id: 4661917b-d50d-4c20-b5e8-8599152e0ca3
Faulting package full name:
Faulting package-relative application ID:Windows 11, CP.AI 2.6.2, ALPR 3.1.0, latest Nvidia Drivers installed on RTX 3090 GPU, the works.
Using YOLOv8 for Object Detection.
ALPR data:
Code:
Module 'License Plate Reader' 3.1.0 (ID: ALPR)
Valid: True
Module Path: <root>\modules\ALPR
AutoStart: True
Queue: alpr_queue
Runtime: python3.9
Runtime Loc: Local
FilePath: ALPR_adapter.py
Start pause: 3 sec
Parallelism: 0
LogVerbosity:
Platforms: all
GPU Libraries: installed if available
GPU Enabled: enabled
Accelerator:
Half Precis.: enable
Environment Variables
AUTO_PLATE_ROTATE = true
CROPPED_PLATE_DIR = <root>\Server\wwwroot
MIN_COMPUTE_CAPABILITY = 6
MIN_CUDNN_VERSION = 7
OCR_OPTIMAL_CHARACTER_HEIGHT = 60
OCR_OPTIMAL_CHARACTER_WIDTH = 30
OCR_OPTIMIZATION = true
PLATE_CONFIDENCE = 0.7
PLATE_RESCALE_FACTOR = 2
PLATE_ROTATE_DEG = 0
REMOVE_SPACES = False
ROOT_PATH = <root>
SAVE_CROPPED_PLATE = true
Status Data: {
"inferenceDevice": "GPU",
"inferenceLibrary": "CUDA",
"canUseGPU": "true",
"successfulInferences": 1,
"failedInferences": 0,
"numInferences": 1,
"averageInferenceMs": 124.0
}
Started: 06 Apr 2024 7:54:47 PM GTB Standard Time
LastSeen: 06 Apr 2024 8:02:17 PM GTB Standard Time
Status: Started
Requests: 2 (includes status calls)
Installation Log
2024-04-06 18:57:45: Installing CodeProject.AI Analysis Module
2024-04-06 18:57:45: ======================================================================
2024-04-06 18:57:45: CodeProject.AI Installer
2024-04-06 18:57:45: ======================================================================
2024-04-06 18:57:45: 522.0Gb of 857Gb available on
2024-04-06 18:57:45: General CodeProject.AI setup
2024-04-06 18:57:45: Creating Directories...done
2024-04-06 18:57:45: GPU support
2024-04-06 18:57:45: CUDA Present...Yes (CUDA 12.3, cuDNN 8.5)
2024-04-06 18:57:45: ROCm Present...No
2024-04-06 18:57:46: Reading ALPR settings.......done
2024-04-06 18:57:46: Installing module License Plate Reader 3.1.0
2024-04-06 18:57:46: Installing Python 3.9
2024-04-06 18:57:46: Python 3.9 is already installed
2024-04-06 18:57:48: Creating Virtual Environment (Local)...done
2024-04-06 18:57:48: Confirming we have Python 3.9 in our virtual environment...present
2024-04-06 18:57:52: Downloading ALPR models...Expanding...done.
2024-04-06 18:57:52: Copying contents of ocr-en-pp_ocrv4-paddle.zip to paddleocr...done
2024-04-06 18:57:52: Installing Python packages for License Plate Reader
2024-04-06 18:57:52: [0;Installing GPU-enabled libraries: If available
2024-04-06 18:57:53: Ensuring Python package manager (pip) is installed...done
2024-04-06 18:57:56: Ensuring Python package manager (pip) is up to date...done
2024-04-06 18:57:56: Python packages specified by requirements.windows.cuda12.txt
2024-04-06 18:58:31: - Installing PaddlePaddle, Parallel Distributed Deep Learning...(✅ checked) done
2024-04-06 18:58:59: - Installing PaddleOCR, the OCR toolkit based on PaddlePaddle...(✅ checked) done
2024-04-06 18:59:00: - Installing imutils, the image utilities library...(✅ checked) done
2024-04-06 18:59:01: - Installing Pillow, a Python Image Library...Already installed
2024-04-06 18:59:01: - Installing OpenCV, the Computer Vision library for Python...Already installed
2024-04-06 18:59:07: - Installing NumPy, a package for scientific computing...Already installed
2024-04-06 18:59:07: Installing Python packages for the CodeProject.AI Server SDK
2024-04-06 18:59:08: Ensuring Python package manager (pip) is installed...done
2024-04-06 18:59:10: Ensuring Python package manager (pip) is up to date...done
2024-04-06 18:59:10: Python packages specified by requirements.txt
2024-04-06 18:59:10: - Installing Pillow, a Python Image Library...Already installed
2024-04-06 18:59:11: - Installing Charset normalizer...Already installed
2024-04-06 18:59:13: - Installing aiohttp, the Async IO HTTP library...(✅ checked) done
2024-04-06 18:59:15: - Installing aiofiles, the Async IO Files library...(✅ checked) done
2024-04-06 18:59:16: - Installing py-cpuinfo to allow us to query CPU info...(✅ checked) done
2024-04-06 18:59:17: - Installing Requests, the HTTP library...Already installed
2024-04-06 18:59:17: Scanning modulesettings for downloadable models...No models specified
2024-04-06 18:59:17: Executing post-install script for License Plate Reader
2024-04-06 18:59:17: Applying PaddleOCR patch
2024-04-06 18:59:17: 1 file(s) copied.
2024-04-06 18:59:24: Self test: Self-test passed
2024-04-06 18:59:24: Module setup time 00:01:38.32
2024-04-06 18:59:24: Setup complete
2024-04-06 18:59:24: Total setup time 00:01:38.68
Installer exited with code 0I don't know what else to do to troubleshoot it. Needless to say, the module never worked, even when checking known static images with clearly visible license plates.
ALPR never worked. Every time I tried to have it recognize a license plate, it crashed. Previous version was 3.0.2, but as I was saying, no version I tried worked.
CP.AI and Blue Iris always worked, first with YOLOv5, then with YOLOv8.
You can try running install_cuDNN.bat to update cuDNN. After running install_cuDNN.bat reboot your PC.
I have done that. Updated cuDNN, rebooted my PC, tried again, with exactly the same symptoms.
LE: "YOLOv8 / license-plate" model recognizes the license plate with the same image used.
LE: "YOLOv8 / license-plate" model recognizes the license plate with the same image used.
Code:
Faulting application name: python.exe, version: 3.9.6150.1013, time stamp: 0x60d9eb23
Faulting module name: cudnn_ops_infer64_8.dll, version: 6.14.11.11070, time stamp: 0x62e9c2eb
Exception code: 0xc0000409
Fault offset: 0x00000000003298d9
Faulting process id: 0x0x3088
Faulting application start time: 0x0x1DA8854F7D125F7
Faulting application path: C:\Program Files\CodeProject\AI247\runtimes\bin\windows\python39\python.exe
Faulting module path: C:\Program Files\NVIDIA\CUDNN\v8.5\bin\cudnn_ops_infer64_8.dll
Report Id: 0f2ec617-7fdd-460b-83e1-71b853d9db52
Faulting package full name:
Faulting package-relative application ID:I notes that the below path is not the default path, it should be just AI. When installing code project did you change the install path?
I had a version prior to AI 2.4.7 which had some issues, and when 2.4.7 came out, I changed the destination directory, because the "AI" one wouldn't work any more, CP.AI was complaining about being unable to install it there.
Since 2.4.7 I used the same folder.
Does ALPR require an absolute path to work?
Since 2.4.7 I used the same folder.
Does ALPR require an absolute path to work?
It should not, but to be sure maybe try uninstall, delete the folders, then reinstall with the default directory.
Installed A2000 into poweredge server with dual E5-2680v3. ALPR CUDA working. CPU mode worked too before I resolved CUDNN issues.
(Used CUDA 12.3 + CUDNN 9. Path to CUDNN bin folder didn't work - duplicated cudnn64_9.dll into new cudnn64_8.dll + put all CUDNN files into C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.3 \ bin + include + lib)
So that confirms it for me - gotta be an issue with old Nehalem CPUs. Good luck.
Well, that solved nothing.
I completely removed CP.AI, deleted all folders, installed from scratch, installed cuDNN, rebooted, started everything up again, the behavior is exactly the same.
Code:
Faulting application name: python.exe, version: 3.9.6150.1013, time stamp: 0x60d9eb23
Faulting module name: cudnn_ops_infer64_8.dll, version: 6.14.11.11070, time stamp: 0x62e9c2eb
Exception code: 0xc0000409
Fault offset: 0x00000000003298d9
Faulting process id: 0x0x61B4
Faulting application start time: 0x0x1DA88E708003D55
Faulting application path: C:\Program Files\CodeProject\AI\runtimes\bin\windows\python39\python.exe
Faulting module path: C:\Program Files\NVIDIA\CUDNN\v8.5\bin\cudnn_ops_infer64_8.dll
Report Id: 02ad5aeb-216f-4cc9-bda1-b62672a751f0
Faulting package full name:
Faulting package-relative application ID:Any idea what to do next?
I have cuDNN 8.5, 8.9 and 9.0 installed, is there a way to use newer cuDNN?
open a command prompt and run set then post the results. This will list all you environment variables.
Is anyone running this with no issues whatsoever? Like months of 24/7 operation with no reboots, crashes etc? Because it seems looking from the outside that it is hard to make it work and not stable at all.
There you go.
Code:
ALLUSERSPROFILE=C:\ProgramData
APPDATA=C:\Users\Eduard Baniceru\AppData\Roaming
CommonProgramFiles=C:\Program Files\Common Files
CommonProgramFiles(x86)=C:\Program Files (x86)\Common Files
CommonProgramW6432=C:\Program Files\Common Files
COMPUTERNAME=WAR-NEW
ComSpec=C:\Windows\system32\cmd.exe
CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.3
CUDA_PATH_V12_3=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.3
DriverData=C:\Windows\System32\Drivers\DriverData
EFC_11900=1
ESET_OPTIONS=
HOMEDRIVE=C:
HOMEPATH=\Users\Eduard Baniceru
IGCCSVC_DB=AQAAANCMnd8BFdERjHoAwE/Cl+sBAAAAu7eR1S9gj0e/Bf61vWzq8QQAAAACAAAAAAAQZgAAAAEAACAAAADAxxbC88w+fwA1tZ5YNvWl92IFOCJJd7mEqjRynwoorAAAAAAOgAAAAAIAACAAAADw46/Jwid6b3w7r6M+SWF5u5VmvGUtSMYsvqf4EnmvyGAAAABzAtZR9f/HCQTNonTRyWI/0e+kTi/Ueqz06xY244G5dBR69GmchGeuaQlkskjAJQkRwwpPmOskqsOZdJXyLzONxiHTsRtsU6Tugoh3JdPxp885ivreL/YutazuTUweNw9AAAAApQPMnXKx9gFpDMNTz8SGYEoJ/9elnuJuIDa4Z/iA69C8wIb86jI5sPGiCMrVFTMML7QP0njvziIeo2z0tz931w==
LOCALAPPDATA=C:\Users\Eduard Baniceru\AppData\Local
LOGONSERVER=\\WAR-NEW
Medit_CHITUBOX_Basic_Bridge=C:\Program Files\CHITUBOX V1.9.5
MOZ_PLUGIN_PATH=C:\Program Files (x86)\Foxit Software\Foxit PDF Reader\plugins\
MSMPI_BENCHMARKS=C:\Program Files\Microsoft MPI\Benchmarks\
MSMPI_BIN=C:\Program Files\Microsoft MPI\Bin\
NUMBER_OF_PROCESSORS=32
OneDrive=C:\Users\Eduard Baniceru\OneDrive
OneDriveConsumer=C:\Users\Eduard Baniceru\OneDrive
OS=Windows_NT
Path=C:\Program Files\NVIDIA\CUDNN\v9.0\bin;C:\Program Files\Microsoft MPI\Bin\;C:\Program Files (x86)\Common Files\Oracle\Java\javapath;C:\Program Files (x86)\VMware\VMware Player\bin\;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.3\bin;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.3\libnvvp;C:\Windows\system32;C:\Windows;C:\Windows\System32\Wbem;C:\Windows\System32\WindowsPowerShell\v1.0\;C:\Windows\System32\OpenSSH\;C:\Program Files\PuTTY\;C:\Program Files\dotnet\;C:\Program Files (x86)\NVIDIA Corporation\PhysX\Common;C:\Program Files\NVIDIA Corporation\Nsight Compute 2023.3.0\;C:\Users\Eduard Baniceru\AppData\Local\Microsoft\WindowsApps;C:\Users\Eduard Baniceru\.dotnet\tools;C:\Program Files\NVIDIA\CUDNN\v8.5\zlib\dll_x64;C:\Program Files\NVIDIA\CUDNN\v8.5\bin;C:\Program Files\Process Lasso\;;f:\AI\CharRTX\\env_nvd_rag\Lib\site-packages\torch\lib;C:\Program Files\NVIDIA Corporation\NVIDIA App\NvDLISR;;C:\Program Files\NVIDIA\CUDNN\v8.9\zlib\dll_x64;C:\Program Files\NVIDIA\CUDNN\v8.9\bin;;C:\Program Files\NVIDIA\CUDNN\v8.9\zlib\dll_x64;C:\Program Files\NVIDIA\CUDNN\v8.9\bin;C:\Users\Eduard Baniceru\AppData\Local\Microsoft\WindowsApps;C:\Users\Eduard Baniceru\.dotnet\tools
PATHEXT=.COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC
PROCESSOR_ARCHITECTURE=AMD64
PROCESSOR_IDENTIFIER=Intel64 Family 6 Model 183 Stepping 1, GenuineIntel
PROCESSOR_LEVEL=6
PROCESSOR_REVISION=b701
ProgramData=C:\ProgramData
ProgramFiles=C:\Program Files
ProgramFiles(x86)=C:\Program Files (x86)
ProgramW6432=C:\Program Files
PROMPT=$P$G
PSModulePath=C:\Program Files\WindowsPowerShell\Modules;C:\Windows\system32\WindowsPowerShell\v1.0\Modules
PUBLIC=C:\Users\Public
SESSIONNAME=Console
SystemDrive=C:
SystemRoot=C:\Windows
TEMP=C:\Users\EDUARD~1\AppData\Local\Temp
TMP=C:\Users\EDUARD~1\AppData\Local\Temp
USERDOMAIN=WAR-NEW
USERDOMAIN_ROAMINGPROFILE=WAR-NEW
USERNAME=Eduard Baniceru
USERPROFILE=C:\Users\Eduard Baniceru
windir=C:\Windows
ZES_ENABLE_SYSMAN=1@MarioBoo I've been using CP.AI for over a year. It's stable as far as I can see.
Started with v2.0.x, now on v2.6.2.
10 cams using CP.AI for person/vehicle/animal detection. 1 cam for ALPR.
Installed using Windows installer on Win10 machine. Most object detection < 100ms due to Nvidia card.
Same machine as BI5 and Home Assistant. All confirmed AI events sent to Home Assistant to provide triggers for automations.
Last edited:
CP.AI itself is stable, but sometimes when upgrading it has the chance to have an issue. Other than that, mine is very stable, except for the ALPR module, which for some reason would crash on successful attempt to read a plate.