Object detection on Nano with yolov8 model

Robotics & Edge Computing Jetson & Embedded Systems Jetson Nano
1

Hello everyone,

Recently I hired someone on Freelancer, to teach a custom model, which I could use for object detection.
He provided me with Yolov8, Onnx and Tensorlite model. Probably Onnx is converted from Pytorch.

I was able to run the inference with Yolov8 and Ultralytics. However the speed was extremely slow (up to 8 FPS). I was never able to run it on GPU. And according to this makes sense.

Could anyone help me to get the inference faster? I need at least 20FPS.
Can it be done with Onnx?

I’m really a beginner and would appreciate if you could point me to an example.

Thank you and kind regards

3

Hi,

On Jetson Nano, YOLOv5s or YOLOv5n can reach > 30fps.

For YOLOv8, you might need a newer device since Nano can only use JetPack 4 and Python 3.6.

Thanks.

5

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.