Updated to the drivers and CUDA versions you were using and got the GPU to be recognized, thanks...
CodeProject.AI Version 2.0
- Thread starter MikeLud1
- Start date
Vettester
Getting comfortable
- Joined
- Feb 5, 2017
- Messages
- 748
- Reaction score
- 704
Hmmm... I guess I'm not sure. I do know that I have two PC's running and had to enable the LPR module on both machines to get it to process the ALPR images. From what I have read the mesh option is a benefit for those that are not using an external GPU and helps with load balancing.
Vettester
Getting comfortable
- Joined
- Feb 5, 2017
- Messages
- 748
- Reaction score
- 704
Yes
You don't HAVE to leave them on. I shut down my slave on the mesh every night, and the BI machine just goes back to its own mode. When I start the slave machine up every morning, then the BI sees its on and starts the mesh with no need for intervention. Mostly at night I have much fewer detections as well.
I have my main machine with Blue Iris/CodeProject Ai running and a 2nd machine running with CodeProject Ai. When I added a third machine with CodeProject Ai it recognizes the other two machines and shows them but there is no activity on the 3rd machine. Just wondering how many machines you can add with the Mesh.
BORIStheBLADE
Getting comfortable
I have read a lot about needing higher end GPUs but I haven't seen this one use many resources when the AI is running. Temp only goes up 2-3 degrees as well..
BORIStheBLADE
Getting comfortable
Still looking for ideas on this error.
Codeproject kept crashing so I deleted it and all files in the codeproject directory and reinstalled Codeproject Ai 2.4.6-Beta. Now I am getting this error, and the ALPR isn't working.
The below is the error I am getting.
15:51:22:ALPR_adapter.py: Traceback (most recent call last):
15:51:22:ALPR_adapter.py: File "C:\Program Files\CodeProject\AI\modules\ALPR\ALPR_adapter.py", line 7, in
15:51:22:ALPR_adapter.py: from request_data import RequestData
15:51:22:ALPR_adapter.py: File "../../SDK/Python\request_data.py", line 7, in
15:51:22:ALPR_adapter.py: from PIL import Image
15:51:22:ALPR_adapter.py: ModuleNotFoundError: No module named 'PIL'
I have no modules named PIL nor have I had any in previous versions of Codeproject Ai.
Any ideas or suggestions??
The below is the error I am getting.
15:51:22:ALPR_adapter.py: Traceback (most recent call last):
15:51:22:ALPR_adapter.py: File "C:\Program Files\CodeProject\AI\modules\ALPR\ALPR_adapter.py", line 7, in
15:51:22:ALPR_adapter.py: from request_data import RequestData
15:51:22:ALPR_adapter.py: File "../../SDK/Python\request_data.py", line 7, in
15:51:22:ALPR_adapter.py: from PIL import Image
15:51:22:ALPR_adapter.py: ModuleNotFoundError: No module named 'PIL'
I have no modules named PIL nor have I had any in previous versions of Codeproject Ai.
Any ideas or suggestions??
MikeLud1
IPCT Contributor
Try uninstalling the ALPR module then enable Do not use download cache before reinstalling the ALPR module
That worked, Thank you Mike!!
- Joined
- Jun 1, 2016
- Messages
- 79
- Reaction score
- 39
This post is long overdue. I'm grateful for the community and hope that there is any easish fix here.
I have 13 cameras. They are all Dauha. All but 1 are IPC-T5442T-ZE.
Running BI 5.8.2.5.
The goal is to send an alert when someone enters the property. With photo if possible.
Camera settings in BI->Trigger tab, using ONVIF/camera events with IVS on most cameras cameras because it seemed more reliable.
Using motion sensor on others and both on a couple.
In the AI tab, if I select use custom models, clicking the three dots doesn't open anything or allow me to select or type anything. It's been a while since I set this up so I'm not recalling how I did it initically.
Issues I'm having:
After the upgrade BI is showing no alerts. Before the upgrade, it was limping along. It's been a long time since it was reliable.
Alert Canceled Nothing found is probably the biggest problem I'm having, although there are several issues. I'd love to get the AI working first.
I turned Facial Recognition off because it repeatedly kept throwing a No Face Found In Image.
On another note, I have a Synology 1621+ so could run docker there. But alas, one thing at a time. Any help would be greatly appreciated. HAPPY HOLIDAYS.
I have 13 cameras. They are all Dauha. All but 1 are IPC-T5442T-ZE.
Running BI 5.8.2.5.
The goal is to send an alert when someone enters the property. With photo if possible.
Camera settings in BI->Trigger tab, using ONVIF/camera events with IVS on most cameras cameras because it seemed more reliable.
Using motion sensor on others and both on a couple.
In the AI tab, if I select use custom models, clicking the three dots doesn't open anything or allow me to select or type anything. It's been a while since I set this up so I'm not recalling how I did it initically.
Issues I'm having:
After the upgrade BI is showing no alerts. Before the upgrade, it was limping along. It's been a long time since it was reliable.
Alert Canceled Nothing found is probably the biggest problem I'm having, although there are several issues. I'd love to get the AI working first.
I turned Facial Recognition off because it repeatedly kept throwing a No Face Found In Image.
On another note, I have a Synology 1621+ so could run docker there. But alas, one thing at a time. Any help would be greatly appreciated. HAPPY HOLIDAYS.
Attachments
My next issue! On my 2nd computer I am running Codeproject MESH. I am getting this error message -
16:16:54:ALPR_adapter.py: W1223 16:16:54.692090 16680 dynamic_loader.cc:274] Note: [Recommend] copy cudnn into CUDA installation directory.
16:16:54:ALPR_adapter.py: For instance, download cudnn-10.0-windows10-x64-v7.6.5.32.zip from NVIDIA's official website,
16:16:54:ALPR_adapter.py: then, unzip it and copy it into C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0
16:16:54:ALPR_adapter.py: You should do this according to your CUDA installation directory and CUDNN version.
I did download cudnn-10.0-windows10-x64-v7.6.5.32.zip and do as the above says but to no avail.
Object detection is running great, but no LPR on the 2nd machine.
16:16:54:ALPR_adapter.py: W1223 16:16:54.692090 16680 dynamic_loader.cc:274] Note: [Recommend] copy cudnn into CUDA installation directory.
16:16:54:ALPR_adapter.py: For instance, download cudnn-10.0-windows10-x64-v7.6.5.32.zip from NVIDIA's official website,
16:16:54:ALPR_adapter.py: then, unzip it and copy it into C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0
16:16:54:ALPR_adapter.py: You should do this according to your CUDA installation directory and CUDNN version.
I did download cudnn-10.0-windows10-x64-v7.6.5.32.zip and do as the above says but to no avail.
Object detection is running great, but no LPR on the 2nd machine.
If you are using camera AI to trigger BI, then CodeProject isn't needed - is that what you want or you want to run the camera ONVIF trigger thru CodeProject?
MikeLud1
IPCT Contributor
What GPU do you have? From what you posted it look like an Nvidia GPU and you do not have CUDA installed.
I downloaded the latest Nvidia GeForce GTX 960 software and installed it to no avail. Still getting the error message.
Rick
MikeLud1
IPCT Contributor
Your GPU is old and is not supported at this time, below are the supported GPUs for the ALPR module
- Joined
- Jun 1, 2016
- Messages
- 79
- Reaction score
- 39
Thanks @wittaj. I wish I knew what was better. I have IVS on some cameras because I thought it was more reliable. But if I can reduce false positives by confirming a human, I think it would be better.
Anything else I can check? Thank you so much for your help.
What should I use for custom models?
Currently using YOLO 5 6.2
Attachments
Last edited: