CodeProject.AI Version 2.5
- Thread starter MikeLud1
- Start date
Here's another set of test images from @Wattstopper101 's cat image with various models:
Same applies as before: Coral TPU, etc. This test image is more interesting because the default COCO label set should identify the cat and car(s). But the cat is viewed from the top down and the cars are very small, so it's a challenge. The only models that can do this reliably are YOLOv5 large, YOLOv8 medium, & large. The small YOLO models are just a bit too small. It also demonstrates the improvement in accuracy of the newer YOLO models.
Same applies as before: Coral TPU, etc. This test image is more interesting because the default COCO label set should identify the cat and car(s). But the cat is viewed from the top down and the cars are very small, so it's a challenge. The only models that can do this reliably are YOLOv5 large, YOLOv8 medium, & large. The small YOLO models are just a bit too small. It also demonstrates the improvement in accuracy of the newer YOLO models.
Last edited:
Here's @tigerwillow1 's night cat scene from page 6. No reliable results, but no false positives in the YOLO models. Sometimes the larger models see the cat, sometimes they don't. More support for the medium YOLOv8 model.
The other images posted to page 6 didn't give any reliable results.
The other images posted to page 6 didn't give any reliable results.
Dudleydogg
Young grasshopper
Dudleydogg
Young grasshopper
Hey.
Anyone noticed YOLOv5.NET 1.9.3 stating cpu being used with a gpu enabled? I know the detection times are definitely the gpu doing the work. Maybe a failed 2.5.4 install.
Anyone noticed YOLOv5.NET 1.9.3 stating cpu being used with a gpu enabled? I know the detection times are definitely the gpu doing the work. Maybe a failed 2.5.4 install.
I have the same. It starts up with CPU with the ObjectDetectionYOLOv5Net process using CPU, it then reports that it lost contact & comes back using GPU for the ObjectDetectionYOLOv5Net process. During this startup phase, BI reports AI not responding. Usually takes about 5 minutes to sort itself out and then runs without issues
im on the 2.5.4 also and it lost contact around 11pm last night...i had to reboot for it to link back up I also noticed that my cuda and cudnn info is not on the sys info screen anymore...i may try a reinstall later
Make sure you have Auto start/stop with Blue Iris disabled. Also refresh the Dashboard web page by doing a CTRL+F5
Make sure you have Auto start/stop with Blue Iris disabled. Also refresh the Dashboard web page by doing a CTRL+F5
I will double check that but I am pretty sure it is unchecked and ctrl f5 refreshed but went to the host denied screen. It is a new install on windows 10 vm with server 2016 host os. There maybe a configuration issue on my end
Dudleydogg
Young grasshopper
@Dudleydogg Shared in ALPR thread: Blue Iris and CodeProject.AI ALPR
License plate #'s are stored in plate field, and also memo field if BI5 AI is configured correctly.
You can search for a plate # in UI3. You can retrieve data via BI5 JSON API, or push data to other systems like Home Assistant via MQTT.
License plate #'s are stored in plate field, and also memo field if BI5 AI is configured correctly.
You can search for a plate # in UI3. You can retrieve data via BI5 JSON API, or push data to other systems like Home Assistant via MQTT.
Last edited:
Dudleydogg
Young grasshopper
I see the tag on the thumbnail but MMO just has car info, something I need to setup somewhere?
where is "Plate Field"
@Dudleydogg Is "Z17C 9RP" the license plate # in your screenshot? That is stored in plate field. You tell BI5 to put it in memo field also if you configure camera AI as suggested by MikeLud1. He's shared his config many times in the ALPR thread. Blue Iris and CodeProject.AI ALPR