Nvidia Jetson Minimalist Images

Motivation

The need for the minimalist images came from the official jetson images being large in size and containing pre-installed packages that are not necessary, resulting in the consumption of valuable disk space and memory.

Supported boards

Spec

OS: ubuntu 20.04

L4T: BSP 32.7.4 & BSP 35.4.1

Download the prebuilt images

The prebuilt images are available in the GitHub release page 👉 here

To be able to decompress the images, you must have lrzip installed.

After the image is decompressed, follow the section below to flash the image to your board.

The default login is:

username: jetson
password: jetson

Build the jetson image yourself

ℹ️ Building the jetson image has been tested only on Linux machines.

Building the jetson image is fairly easy. All you need to have is the following tools installed on your machine.

Start by cloning the repo from github

$ git clone https://github.com/pythops/jetson-image
$ cd jetson-image

Then create a new rootfs using the following command:

$ just build-jetson-rootfs

This will create the rootfs in the rootfs directory.

ℹ️ You can modify the Containerfile.rootfs file to add any tool or configuration that you will need in the final image.

Next, use the following command to build the Jetson image:

$ just build-jetson-image -b <board> -r <revision> -d <device>

For example, to build an image for jetson-orin-nano board:

$ just build-jetson-image -b jetson-orin-nano -d SD

Run with -h for more information

$ just build-jetson-image -h

The Jetson image will be built and saved in the current directory in a file named jetson.img

Flashing the image into your board

To flash the jetson image, just run the following command:

$ sudo just flash-image <jetson image file> <device>

Where device is the name of the sdcard/usb identified by your system. For instance, if your sdard is recognized as /dev/sda, then replace device by /dev/sda

ℹ️ There are numerous tools out there to flash images to sd card that you can use. I stick with dd as it's simple and does the job

Nvidia Libraries

Once you boot the board with the new image, then you can install Nvidia libraries using apt

$ sudo apt install -y libcudnn8 libcudnn8-dev ...

Result

For the jetson nano for instance with the new image, only 150MB of RAM is used, which leaves you with 3.85 GB for your projects !

If you find this helpful, don't forget to give it a star ⭐

Looking for professional support ?

If you need more advanced configuration or a custom setup, you can contact me on this address support@pythops.com

License

AGPLv3

Gurkengewuerz/jetson-image