Drone Detection using YOLOv8

Robotics & Edge Computing Jetson Systems Jetson Orin Nano
,
1

Hii,
I am developing a project on drone detection using Jetson orin nano developer kit 8GB ram and I have also done it using YOLOv8n model but when I switch to YOLOv8m model the FPS becomes very slow means video is coming on live detection after 5-6 sec so I am trying it to run on GPU enabled CUDA and I try to install GPU enabled pytorch and torchvision but it fails every time and install CPU enable pytorch and torchvision so how to install it

  1. by this should i get rid of from low FPS problem
  2. Sometime it also gives false detection of drone as I have trained it on only drone based dataset so how to solve these problem as I have to work for professional use
6

Hi,

Please install related prebuilt wheels from jp6/cu126 index and check whether issue occurs.

Thanks

7

I did the same project, never install anything manual, use the SDK manager and let it handle everything for you unless you want to enter a world full of bugs, I made it works manually but it will break with the first update.

8

You can find PyTorch and Torchvision installation guide in Ultralytics Docs