Deploy Computer Vision Models on Edge Devices such as NVIDIA Jetson Orin Nano

Learning Objectives

At the end of this theory note, trainees must be able to:

1. Set up and configure a Jetson Orin Nano with JetPack, drivers, and required libraries.
2. Convert and optimise a trained computer vision model for edge inference within hardware constraints.
3. Connect cameras and sensors to an edge device and execute a deployed model for real-time inference.
4. Benchmark and evaluate edge inference performance.
5. Diagnose and troubleshoot common edge deployment issues.
6. Document an edge deployment in accordance with reproducibility standards.