New CodeProject.AI Object Detection (YOLO11 .NET) Module
- Thread starter MikeLud1
- Start date
@MikeLud1 - Huge thanks for continuing Code Project AI development. Yolo11.net just works, and works well for me so far. Using ipcam-combined-v5 model in BI 6.0.1 results in practically no missed detections and really low inference times just using CPU on my test build.
Initially, I was getting 400 -700ms detection times, until I switched to a custom model, and now I get this:
Thanks again Mike, and Happy New Year!
Initially, I was getting 400 -700ms detection times, until I switched to a custom model, and now I get this:
Thanks again Mike, and Happy New Year!
Just curious if its possible to detect errors while 3d printing? A lot of the more expensive printers have cameras built in that detect when something goes wrong while printing. eg. object falls over, a nest ball of filament, clog, etc etc.
So this is only for GPU, no CPU support ?
Currently running with YOLOv5 6.2 (Medium model size) in a Ubuntu LXC in Proxmox on a I5-10500T CPU, and quite happy with it. (80-250ms)
Currently running with YOLOv5 6.2 (Medium model size) in a Ubuntu LXC in Proxmox on a I5-10500T CPU, and quite happy with it. (80-250ms)
It works with CPU only also, mine is running on that as it also is in a vm. I keep bouncing back and forth, I see slightly better times using code project for AI vs the now built in AI models in BI.
neutrinobox
n3wb
hi there, i have a little question. I use yolov8 with a small v8 pva model (person, vehicle, animal) on an old desktop with i7 4790k, 32gb ram, Win10 lite and a gtx960. Yolov8 runs with cuda support , nvdec decodes the video. At all i have good detection around 38....50ms with very low load but i have a lot of false readings (table as person, chair -> animal) at low light and IR light conditions. I think it depends on model, so... make it sense to switch to yolo11 with mikeLud costum models ? I read a lot for cuda.... but works yolo11 with old maxwell hardware ?
Thanks for any answer
Thanks for any answer
Bruce_H
Getting comfortable
Can you tell us what version of BI you are using and are you using CPAI or the built in AI in BlueIris, hard to make any suggestions with out this info!hi there, i have a little question. I use yolov8 with a small v8 pva model (person, vehicle, animal) on an old desktop with i7 4790k, 32gb ram, Win10 lite and a gtx960. Yolov8 runs with cuda support , nvdec decodes the video. At all i have good detection around 38....50ms with very low load but i have a lot of false readings (table as person, chair -> animal) at low light and IR light conditions. I think it depends on model, so... make it sense to switch to yolo11 with mikeLud costum models ? I read a lot for cuda.... but works yolo11 with old maxwell hardware ?
Thanks for any answer