Getting Started with Jetson Orin Nano

This repository is a guide to help you get started with the Jetson Orin Nano Developer Kit. It walks through the essential setup steps, recommended tools, and best practices to prepare your board for AI and embedded systems projects.

The Jetson Orin Nano is a compact AI development board from NVIDIA, designed for edge computing and robotics. It offers powerful GPU and CPU performance, supports deep learning frameworks, and is ideal for running computer vision, machine learning, and embedded AI applications.

Repository Structure

• Docs: Step-by-step setup and usage documentation for the Jetson Orin Nano.
• Examples: Scripts demonstrating how to use the Jetson Orin Nano with other tools.
• Images: Visuals and screenshots related to setup or board usage.

Initial Setup Overview

You can set up the Jetson Orin Nano from scratch using the steps in the first guide. The summary below shows what you'll need and how to get started.

• What you Need Before you Start: Tools and hardware required for setup.
• Installing the Operating System: Steps to flash the Jetson OS image onto a microSD card.
• Adding NVMe SSD Storage: How to mount an SSD to extend the board's storage.
• Powering Up and First Boot: Connecting peripherals and booting into the OS for the first time.

Running a Local LLM

You can run a language model locally on the Jetson Orin Nano using Ollama and Open WebUI. The steps below show how to install and get them running.

Installing Ollama

git clone https://github.com/dusty-nv/jetson-containers

bash jetson-containers/install.sh

jetson-containers run --name ollama $(autotag ollama)

Running an LLM on the Command Line

ollama run llama3.2:3b

To end the session, type:

/bye

Running an LLM on Open WebUI

sudo docker run -d --network=host \
    -v ${HOME}/open-webui:/app/backend/data \
    -e OLLAMA_BASE_URL=http://127.0.0.1:11434 \
    --name open-webui \
    --restart always \
    ghcr.io/open-webui/open-webui:main

Then open a browser and go to: http://JETSON_IP:8080

Running a Remote Local AI Server

The third guide in the Docs folder shows how to access your Jetson Orin Nano remotely from another computer on the same network. This allows you to treat the Jetson as a local AI server.

Running Stable Diffusion

The fourth guide in the Docs folder explains how to run Stable Diffusion locally on the Jetson Orin Nano using jetson-containers.

Run one of the following commands:

jetson-containers run $(autotag stable-diffusion-webui)

or

jetson-containers run dustynv/stable-diffusion-webui:r35.3.1

Once it starts, check the terminal output for the local URL to access the Stable Diffusion WebUI.

Integrating a Remote MySQL Database

You can configure and connect a MySQL database to the Jetson Orin Nano for data management and application integration.

MySQL Server Setup on Ubuntu 22.04

sudo apt update

sudo apt-get install mysql-server

systemctl is-active mysql

sudo mysql_secure_installation

Then log in:

sudo mysql

ALTER USER 'root'@'localhost' IDENTIFIED WITH mysql_native_password BY 'Password123#@!';
FLUSH PRIVILEGES;
QUIT

And reconnect:

mysql -u root -p

Install MySQL Workbench

sudo snap install mysql-workbench-community

Then launch it from the Applications menu.

Create and Use a Database

mysql -u root -p
CREATE DATABASE database_name;
SHOW DATABASES;
USE database_name;

Connect the Jetson to a Remote MySQL Server

Edit MySQL config:

sudo nano /etc/mysql/mysql.conf.d/mysqld.cnf

Change bind-address = 127.0.0.1 to bind-address = 0.0.0.0

Restart MySQL:

sudo systemctl restart mysql

Allow remote access:

sudo mysql
GRANT ALL PRIVILEGES ON *.* TO 'root'@'%' IDENTIFIED BY 'Password123#@!';
FLUSH PRIVILEGES;

Install Python connector:

pip install mysql-connector-python

Use Python to connect using IP, credentials, and query the database. You can find example scripts for this inside the Examples folder.

Repository Goals

To make it easy for beginners and first-time users to:

• Set up and boot the Jetson Orin Nano successfully
• Understand the board’s basic interfaces and boot process
• Prepare the device for AI/robotics applications

Related Work

This repository is part of my PFA (Projet de Fin d'Année). The project is about smart cities and the use of smart sensors to count vehicles and measure things like traffic flow, speed, pollution, and more. The Jetson Orin Nano is used as the main controller. It will be linked to the main project repository once completed.