Thanks, not sure what steps to take to fix the falling back to single-TPU issue. I have completely re-installed everything from scratch including creating a new docker image for CPAI. Should I revert to an older version?
CodeProject.AI Version 2.5
- Thread starter MikeLud1
- Start date
Yeah, I’m afraid that I don’t have much advice for you. It’s been a few months since I looked at the code. The single-TPU fallback would need to be debugged, but I’m not confident what that means or how to do it here. Different versions would be a good way to get things working more easily.
Village Guy
Pulling my weight
The CodeProject team is current redoing the website where CodeProject.AI is stored bear with them until they sort it all out.
I recommend staying on the current version you have install until they sort it all out.
New Discussions Link (moved to GitHub)
https://github.com/codeproject/CodeProj ... iscussions
Download link (currently still under development file downloads do not work)
CodeProject.AI Server - CodeProject.AI Server v2.8.1
Documentation
CodeProject.AI Server - CodeProject.AI Server v2.9.0
I recommend staying on the current version you have install until they sort it all out.
New Discussions Link (moved to GitHub)
https://github.com/codeproject/CodeProj ... iscussions
Download link (currently still under development file downloads do not work)
CodeProject.AI Server - CodeProject.AI Server v2.8.1
Documentation
CodeProject.AI Server - CodeProject.AI Server v2.9.0
Village Guy
Pulling my weight
I already have a YOLO11 ipcam-combined model with @algertc help training the model. It will not work with the YOLOv8 module. I have the YOLO11 module ready to release, I will see if Chris from CP.AI will release the module.
GitHub - MikeLud/CodeProject.AI-Custom-IPcam-Models
Contribute to MikeLud/CodeProject.AI-Custom-IPcam-Models development by creating an account on GitHub.
github.com
The plan is to release all my custom models.
In that case, I'm going to start training a YOLO11 version of my delivery model.
GitHub - VideoDad/AI-Models: Custom Trained AI Models for object detection
Custom Trained AI Models for object detection. Contribute to VideoDad/AI-Models development by creating an account on GitHub.
github.com
slidermike
Getting the hang of it
I implemented a custom DeepStack model to alert me when the mail is delivered because my mailbox is down the street from my home. I'm using a camera mounted near the street that also detects people and vehicles (including ALPR).
The post truck detection works but will fail if the event is preceded by another event. For example, a jogger triggered an alert at 11:18:29 which then masked the mail delivery 6 seconds later at 11:18:35. Is there a way to ensure sequential alerts fire?
The post truck detection works but will fail if the event is preceded by another event. For example, a jogger triggered an alert at 11:18:29 which then masked the mail delivery 6 seconds later at 11:18:35. Is there a way to ensure sequential alerts fire?
I've added details to the readme page of the GitHub repository.
GitHub - VideoDad/AI-Models: Custom Trained AI Models for object detection
Custom Trained AI Models for object detection. Contribute to VideoDad/AI-Models development by creating an account on GitHub.
github.com
Let me know if there are any steps I can clarify further.
I have to say thank you again @VideoDad, I really enjoyed using the delivery model. Unfortunately, where we live now we rarely get Home Deliveries, even though I still run the model. But, actually fortunately we are away from the big city and enjoying our retirement out in the country...I've added details to the readme page of the GitHub repository.
GitHub - VideoDad/AI-Models: Custom Trained AI Models for object detection
Custom Trained AI Models for object detection. Contribute to VideoDad/AI-Models development by creating an account on GitHub.
Let me know if there are any steps I can clarify further.
Also, thank you @MikeLud1 for your hard work too. You both Rock!!!
slidermike
Getting the hang of it
Thank you very much VD.I've added details to the readme page of the GitHub repository.
GitHub - VideoDad/AI-Models: Custom Trained AI Models for object detection
Custom Trained AI Models for object detection. Contribute to VideoDad/AI-Models development by creating an account on GitHub.
Let me know if there are any steps I can clarify further.
That was super helpful.
I do have one question.
On your github page the readme says to use a field called "Mark as Vehicle" which I do not appear to have in my BI.
I am running the most recent, stable branch with current CPIA.
The rest of your easy to follow directions were accurate but I am curious how I am not seeing this field.
Camera Settings>Alert tab>AI Configuration>
If I read your directions correctly, I should see that field to make as a vehicle but I do not.
In the latest version it's a global AI setting. So go to the main settings page > AI. Add it to the list of objects that are already considered vehicles.
The YOLO11 version of the delivery model (delivery.pt) has been trained and added to my GitHub repository.In that case, I'm going to start training a YOLO11 version of my delivery model.
GitHub - VideoDad/AI-Models: Custom Trained AI Models for object detection
Custom Trained AI Models for object detection. Contribute to VideoDad/AI-Models development by creating an account on GitHub.
github.com
As @MikeLud1 mentioned, he's working with ChrisMaunder to get YOLO11 added as an installable module for CPAI shortly.
YOLO v11 · codeproject CodeProject.AI-Server · Discussion #286
Hello, Any chance to integrate new YOLO versions ? Thanks.
github.com
I've also tried to get Chris Maunder to include my delivery custom models pre-installed in each of the CPAI modules (YOLOv5 6.2, YOLOv5 .NET, YOLOv8, etc.) but so far, no response. I hope this can happen with the next updates so that it won't be an additional manual step for people.
Last edited:
If we only could have a model for Solicitors/Salespeople LOLTHe YOLO11 version of the delivery model (delivery.pt) has been trained and added to my GitHub repository.
GitHub - VideoDad/AI-Models: Custom Trained AI Models for object detection
Custom Trained AI Models for object detection. Contribute to VideoDad/AI-Models development by creating an account on GitHub.
As @MikeLud1 mentioned, he's working with ChrisMaunder to get YOLO11 added as an installable module for CPAI shortly.
YOLO v11 · codeproject CodeProject.AI-Server · Discussion #286
Hello, Any chance to integrate new YOLO versions ? Thanks.
I've also tried to get Chris Maunder to include my delivery custom models pre-installed in each of the CPAI modules (YOLOv5 6.2, YOLOv5 .NET, YOLOv8, etc.) but so far, no response. I hope this can happen with the next updates so that it won't be an additional manual step for people.
Vettester
Getting comfortable
- Feb 5, 2017
- 916
- 894
