|
Hi,
Im using Blue Iris in a Win Srv 2019 HyperV-VM on a Dell Poweredge T320, Xeon 2470, 16GB RAM with a Quado M2000 passed through via DDA.
I had 1.5 up and running but CPU is a bottleneck and was waiting for GPU support to be released, but i need some very basic help with the install.
Ive installed Nvidia 516.94 (516.94-nvidia-rtx-winserv-2016-2019-2022-64bit-international-dch-whql.exe)
Ive installed the CUDA 11.7 Toolkit
Ive downloaded the script
First question, its refering to: "1. we're in the source code /Installers/ directory."
Which installer directory?
|
|
|
|
|
I just updated to 1.5.6.2 and it will not stay running on my non-GPU Windows 10 system.
Is there somewhere I can download a pre-GPU version so I can have a stable system again? I'm not having any luck finding previous releases.
Below is a full copy/paste of the log on my system. As soon as I restart it it falls on its face again.
ass="debug ">1:38:23 PM: CodeProject.BackendProcessRunner: YOLOv5_VERBOSE = false
1:38:23 PM: CodeProject.BackendProcessRunner: ------------------------------------------------------------------
1:38:23 PM: CodeProject.BackendProcessRunner: Starting C:\Program Files...\CodeProject\AI\Python "C:\Program Files...\CodeProject\AI\scene.py"
1:38:23 PM: CodeProject.BackendProcessRunner: Started Scene Classification backend
1:38:23 PM: CodeProject.BackendProcessRunner: Attempting to start Scene Classification, Runtime: python37, FilePath: Vision\intelligencelayer\scene.py
1:38:23 PM: CodeProject.BackendProcessRunner: Setting Environment variables for Scene Classification
1:38:23 PM: CodeProject.BackendProcessRunner: ------------------------------------------------------------------
1:38:23 PM: CodeProject.BackendProcessRunner: CPAI_APPROOTPATH = C:\Program Files\CodeProject\AI
1:38:23 PM: CodeProject.BackendProcessRunner: CPAI_ERRLOG_APIKEY = ed359c3a-8a77-4f23-8db3-d3eb5fac23d9
1:38:23 PM: CodeProject.BackendProcessRunner: CPAI_PORT = 5000
1:38:23 PM: CodeProject.BackendProcessRunner: APPDIR = C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\intelligencelayer
1:38:23 PM: CodeProject.BackendProcessRunner: DATA_DIR = C:\ProgramData\CodeProject\AI
1:38:23 PM: CodeProject.BackendProcessRunner: MODE = MEDIUM
1:38:23 PM: CodeProject.BackendProcessRunner: MODELS_DIR = C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\assets
1:38:23 PM: CodeProject.BackendProcessRunner: PROFILE = desktop_gpu
1:38:23 PM: CodeProject.BackendProcessRunner: TEMP_PATH = C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\tempstore
1:38:23 PM: CodeProject.BackendProcessRunner: USE_CUDA = True
1:38:23 PM: CodeProject.BackendProcessRunner: VISION-SCENE = True
1:38:23 PM: CodeProject.BackendProcessRunner: YOLOv5_VERBOSE = false
1:38:23 PM: CodeProject.BackendProcessRunner: ------------------------------------------------------------------
1:38:23 PM: CodeProject.BackendProcessRunner: Starting C:\Program Files...\CodeProject\AI\Python "C:\Program Files...\CodeProject\AI\scene.py"
1:38:23 PM: CodeProject.BackendProcessRunner: Started Scene Classification backend
1:38:25 PM: CodeProject.QueueServices: Queued: 'detect' request, id ed5a1fff-72fe-490f-8e9a-209a51d2f4b0
1:38:25 PM: CodeProject.QueueServices: Queued: 'detect' request, id ed5a1fff-72fe-490f-8e9a-209a51d2f4b0
1:38:25 PM: CodeProject.QueueServices: Queued: 'detect' request, id 5a30de19-9b0e-467c-90da-12c78034c18f
1:38:27 PM: CodeProject.QueueServices: Queued: 'detect' request, id b358765e-111e-49a9-b302-db8148108f6b
1:38:27 PM: CodeProject.QueueServices: Queued: 'detect' request, id 808a8613-4844-4d8f-af3e-861efb4b5385
1:38:27 PM: CodeProject.QueueServices: Queued: 'detect' request, id 05f0622b-40c1-4a35-867e-07865566fda3
1:38:27 PM: CodeProject.QueueServices: Queued: 'detect' request, id cb3d7b43-e198-4f53-b10d-2c244ac10034
1:38:27 PM: CodeProject.QueueServices: Queued: 'detect' request, id 863e97a9-f8c3-4415-9af5-7a5587a6168e
1:38:27 PM: CodeProject.QueueServices: Queued: 'detect' request, id 0c507f19-3a9d-4e9a-b789-6314a4e9a631
1:38:27 PM: CodeProject.QueueServices: Queued: 'detect' request, id baab4ac9-ef72-43b5-b5b5-76f71d49713a
1:38:28 PM: CodeProject.BackendProcessRunner: Attempting to start Object Detection (Python), Runtime: python37, FilePath: Vision\intelligencelayer\detection.py
1:38:28 PM: CodeProject.BackendProcessRunner: Setting Environment variables for Object Detection (Python)
1:38:28 PM: CodeProject.BackendProcessRunner: ------------------------------------------------------------------
1:38:28 PM: CodeProject.BackendProcessRunner: CPAI_APPROOTPATH = C:\Program Files\CodeProject\AI
1:38:28 PM: CodeProject.BackendProcessRunner: CPAI_ERRLOG_APIKEY = ed359c3a-8a77-4f23-8db3-d3eb5fac23d9
1:38:28 PM: CodeProject.BackendProcessRunner: CPAI_PORT = 5000
1:38:28 PM: CodeProject.BackendProcessRunner: APPDIR = C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\intelligencelayer
1:38:28 PM: CodeProject.BackendProcessRunner: DATA_DIR = C:\ProgramData\CodeProject\AI
1:38:28 PM: CodeProject.BackendProcessRunner: MODE = MEDIUM
1:38:28 PM: CodeProject.BackendProcessRunner: MODELS_DIR = C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\assets
1:38:28 PM: CodeProject.BackendProcessRunner: PROFILE = desktop_gpu
1:38:28 PM: CodeProject.BackendProcessRunner: TEMP_PATH = C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\tempstore
1:38:28 PM: CodeProject.BackendProcessRunner: USE_CUDA = True
1:38:28 PM: CodeProject.BackendProcessRunner: VISION-DETECTION = True
1:38:28 PM: CodeProject.BackendProcessRunner: YOLOv5_VERBOSE = false
1:38:28 PM: CodeProject.BackendProcessRunner: ------------------------------------------------------------------
1:38:28 PM: CodeProject.BackendProcessRunner: Starting C:\Program Files...\CodeProject\AI\Python "C:\Program Files...\CodeProject\AI\detection.py"
1:38:28 PM: CodeProject.BackendProcessRunner: Started Object Detection (Python) backend
1:38:30 PM: CodeProject.QueueServices: Queued: 'detect' request, id 5ab84fed-ed1f-4fc1-8754-d016c382e17c
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: Vision AI services setup: Retrieving environment variables...
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: Traceback (most recent call last):
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: File "C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\intelligencelayer\detection.py", line 56, in
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: detector = YOLODetector(model_path, reso, cuda=USE_CUDA)
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: File "C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\intelligencelayer\.\process.py", line 21, in __init__
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: self.model = attempt_load(model_path, device=self.device)
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: File "C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\intelligencelayer\.\models\experimental.py", line 82, in attempt_load
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: model.append(ckpt.fuse().eval() if fuse else ckpt.eval()) # fused or un-fused model in eval mode
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: File "C:\Program Files\CodeProject\AI\AnalysisLayer\Vision\intelligencelayer\.\models\yolo.py", line 232, in fuse
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: m.conv = fuse_conv_and_bn(m.conv, m.bn) # update conv
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: File "C:\Program Files\CodeProject\AI\AnalysisLayer\bin\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 802, in __setattr__
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: remove_from(self.__dict__, self._parameters, self._buffers, self._non_persistent_buffers_set)
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: File "C:\Program Files\CodeProject\AI\AnalysisLayer\bin\python37\venv\lib\site-packages\torch\nn\modules\module.py", line 772, in __getattr__
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: type(self).__name__, name))
1:38:31 PM: CodeProject.BackendProcessRunner: detection.py: torch.nn.modules.module.ModuleAttributeError: 'Conv' object has no attribute '_non_persistent_buffers_set'
1:38:33 PM: CodeProject.QueueServices: Queued: 'detect' request, id 31b5419c-3cc4-4568-ad02-1c2839d7f231
|
|
|
|
|
|
I spent several hours chasing my tail on this today but couldn't seem to make any headway. I was unable to find a pre-GPU support Windows package available anywhere for download so I had no choice but to go back to the "Deep end" to get my alerts working again.
Edit: I ended up abandoning DS and trying AI Server again. Even after a fresh installation 1.5.6.2 insisted on trying to run on my inadequate GPU. I've deleted all traces of anything NVIDIA or CUDA related I can find on my system.
9:02:22 PM: CodeProject.BackendProcessRunner: USE_GPU = True
9:02:27 PM: CodeProject.BackendProcessRunner: PROFILE = desktop_gpu
9:02:27 PM: CodeProject.BackendProcessRunner: USE_CUDA = True
I eventually found a way to get 1.5.5 downloaded again and should (hopefully) be OK on that release until this is sorted. If anyone else finds themselves in a similar situation and needs to go back to a pre-GPU Windows release, you can use this link to get it: https://www.codeproject.com/KB/Articles/5322557/CodeProject.AI.Server-1.5.5.zip
|
|
|
|
|
|
Running 1.5.6.0002 and what a huge improvement. Using BI 5.6.0.2 on W11 I7-6700k. The installer ran perfect.
A big Thank You for your great work !

|
|
|
|
|
That looks great! Think you might be able to provide an update on night time performance later? Is your rig dedicated to BI, or do you have a multipurpose setup? <-- just a general question as maybe its also doing other processing (plex, VMs, etc) I'd like to drop my GPU for my machine in favor of my 9700k.
|
|
|
|
|
I myself do not use AI at night. This is a dedicated pc just for BI and it has 16GB ram.
|
|
|
|
|
I agree, custom models went from 1000ms to 120ms on a ryzen 1700! thank you guys.
|
|
|
|
|
Would be nice if I could get custom models working...
|
|
|
|
|
registry editor on windows, need to check the box for deepstack_custom and fill in the correct custom path
|
|
|
|
|
I'm running AI on a separate Linux server from BI. I've asked about how to get this working but I'm afraid it's been buried at this point.
|
|
|
|
|
To get BI to work in your case create a folder on the BI server C drive "CustomModels" and copy the custom models you want to use to this folder and point BI to this folder.
|
|
|
|
|
At the risk of repeating myself... this does not work and I fail to see how setting BI to use custom models tells AI on a completely separate system to use custom models...
|
|
|
|
|
Did you try it, during my testing I have CodeProject.AI on one PC and Blue Iris on another PC and it works. Post some screenshots of the Blue Iris main AI settings and the folder where the models are saved.
|
|
|
|
|
Ah ok then. No worries. When I get some free time I'll mess around with it to see how the night performance is. I'd imagine it'd be matching or close to matching the existing performance.
|
|
|
|
|
Chris,
What Runtime and versions are needed? Maybe I am out of date with 2019.
I am struggling with some errors in Server 2019 when using a GPU (NVIDIA P620)
The Same Setup & GPU works perfectly with W10 21H2.
Setup:
en-us_windows_server_2019_updated_aug_2021_x64_dvd_a6431a28.iso
NVIDIA Quadro P620
Driver - 516.94-quadro-rtx-desktop-notebook-win10-win11-64bit-international-dch-whql.exe
NVIDIA Toolkit - cuda_11.7.1_516.94_windows.exe
install_CUDnn.bat - using cudnn-windows-x86_64-8.5.0.96_cuda11-archive.zip & zlib123dllx64.zip in Install dir of CodeProject.AI-Server-main.zip - dated 1 week ago.
Again ... Same setup on W10 21H2 works Great!!
I see all Timeouts in Bi with this install on Server 2019 and it generates these errors.
Windows Application Error log
--------------------
EventID 1000
Category: CodeProject.AI.Analysis.Yolo.ObjectDetector
EventId: 0
Unable to run prediction
Exception:
Microsoft.ML.OnnxRuntime.OnnxRuntimeException: [ErrorCode:RuntimeException] Non-zero status code returned while running Mul node. Name:'Mul_2' Status Message: D:\a\_work\1\s\onnxruntime\core\framework\bfc_arena.cc:342 onnxruntime::BFCArena::AllocateRawInternal Failed to allocate memory for requested buffer of size 19660800
at Microsoft.ML.OnnxRuntime.NativeApiStatus.VerifySuccess(IntPtr nativeStatus)
at Microsoft.ML.OnnxRuntime.InferenceSession.RunImpl(RunOptions options, IntPtr[] inputNames, IntPtr[] inputValues, IntPtr[] outputNames, DisposableList`1 cleanupList)
at Microsoft.ML.OnnxRuntime.InferenceSession.Run(IReadOnlyCollection`1 inputs, IReadOnlyCollection`1 outputNames, RunOptions options)
at Microsoft.ML.OnnxRuntime.InferenceSession.Run(IReadOnlyCollection`1 inputs, IReadOnlyCollection`1 outputNames)
at Microsoft.ML.OnnxRuntime.InferenceSession.Run(IReadOnlyCollection`1 inputs)
at Yolov5Net.Scorer.YoloScorer`1.Inference(Image image)
at Yolov5Net.Scorer.YoloScorer`1.Predict(Image image)
at CodeProject.AI.Analysis.Yolo.ObjectDetector.Predict(Byte[] imageData)
---------------
Any guidance would be appreciated....
Thanks,
Cj
|
|
|
|
|
Try the latest version (1.5.6.2). It doesn't use that module for object detection any more.
Also, how much memory does the 2019 server have?
"Time flies like an arrow. Fruit flies like a banana."
modified 17hrs ago.
|
|
|
|
|
Matthew,
Ok .. Will upgrade!!!
Its a Dell Poweredge T420 .. Dual E5-2470v2 with 48GB ram
Thanks,
Cj
|
|
|
|
|
I had a typo in my original reply. The new version is 1.5.6.2
"Time flies like an arrow. Fruit flies like a banana."
|
|
|
|
|
It's now working
BIG Thanks to the Dev Team.....
T420 - Dual E5-2470v2 48GB Ram
Quadro P620
Server 2019 eval
Using CodeProject.Ai 1.5.6.2 and Bi 5.5.8.2 with 15 cameras continuous record
Memory usage is 15.5 GB ram with 32.2gb to spare...
18% avg CPU load with no triggers....
Playing so far I am getting Person recognition in 182ms to 219ms (Non GPU averaged 500ms to 750ms)
My Cat, However is a little upset as he is now recognized as a DOG!! ( I am arranging counseling sessions )
And its all standalone!!!!! No internet connection after install....
The FUN Begins
Cj
|
|
|
|
|
CJ Davis wrote: My Cat, However is a little upset as he is now recognized as a DOG!! ( I am arranging counseling sessions )
That's too funny.
OK Mike - we need a better animal detector!
cheers
Chris Maunder
|
|
|
|
|
Below are the accuracy stats of the animal model, cat did have good results. I am going to start retraining all my models in the upcoming week or two using yolov5 so hopefully I get better results

|
|
|
|
|
Well...
Maybe CJ should post a picture of his cat before deciding this is truly a technical problem. 
|
|
|
|
|
Ok ... Here Ya go...

[
{
"api":"objects",
"found":{
"predictions":[
{
"label":"bicycle",
"confidence" .4944458,
"y_min":1157,
"x_min":4,
"y_max":1627,
"x_max":345}
,
{
"label":"truck",
"confidence" .8355808,
"y_min":603,
"x_min":745,
"y_max":958,
"x_max":1712}
,
{
"label":"car",
"confidence" .6874123,
"y_min":1538,
"x_min":3180,
"y_max":2147,
"x_max":3837}
,
{
"label":"dog",
"confidence" .65638524,
"y_min":1404,
"x_min":457,
"y_max":1740,
"x_max":768}
]
,
"success":true}
}
,
{
"api":"IPcam-animal",
"found":{
"success": true,
"predictions": [
]
}
}
,
{
"api":"IPcam-dark",
"found":{
"success": true,
"predictions": [
{
"confidence": 0.8843992948532104,
"label": "Car",
"x_min": 754,
"y_min": 613,
"x_max": 1878,
"y_max": 981}
]
}
}
]
Now to be fair My Cat 'Essaf' was a little happier about this one:

So I am sure you are wondering what kind of a name Essaf is.
Well He was a feral kitty that was walking down the drive one day with the rest of the litter following Mommy.
He got pushed out when Mom dropped the mouse and they all ate but him.
I told the wife if she got busy and fed that kitty he would stick around and could be getting those mice that get into her car and stink things up!
After the de-worming,de-ticking, de-lice and such ... Well the rest is history except when I took him to the vet the first time.
The very nice Young lady asked "What's his Name?".
I didn't have a name So ... Lets see .. He 'E'ats, 'S'leeps, 'S' does his business 'A'nd lets 'F' Terrible Wind .. Perfect!!
We really get BIG smiles when Essaf and I visit the Vet!!!
Cj
|
|
|
|
|