Skip to main content

Jetson Nano setup

Installation​

  • Prepare SD card
  • Insert SD card
  • Run device
    • Connect wifi antennas
    • Connect camera cables
    • Connect Power, Reset and LED cables in correct pins if you use a case
    • Set pin near the output barrel - this tells device to use 5V power supply
    • Use HDMI output - Display port does not work by default on boot
  • After OS installation is complete, you won’t have wifi right away - you need a restart

Global update​

sudo apt-get -y update
sudo apt-get upgrade

# Uninstall LibreOffice to save space
sudo apt remove --purge libreoffice* -y
sudo apt-get clean -y
sudo apt autoremove -y
sudo apt-get update

# Install curl
sudo apt install curl

# Docker upgrade, use own username
sudo usermod -aG docker gratheon
sudo apt-get --only-upgrade install docker.io

# Add docker-compose
export DOCKER_COMPOSE_VERSION=1.27.4
sudo apt-get install libhdf5-dev
sudo apt-get install libssl-dev
sudo pip3 install docker-compose=="${DOCKER_COMPOSE_VERSION}"

# to not display terminal errors when playing annoying sounds
sudo apt install libcanberra-gtk-module libcanberra-gtk3-module -y

# video cam utils
sudo apt-get install v4l-utils

How to install ML software with GPU acceleration​

Python​

sudo apt install python3-pip
pip3 install --upgrade pip

Opencv with cuda​

Install cuDNN​

https://developer.nvidia.com/cudnn-downloads

Install jtop to see GPU usage in realtime

# update pip as root
sudo curl <https://bootstrap.pypa.io/get-pip.py> -o get-pip.py
sudo python get-pip.py
sudo python -m pip install jetson-stats

# restart needed
jtop
nvidia-smi

Install Pytorch with CUDA​

see https://developer.download.nvidia.com/compute/redist/jp/v60dp/pytorch/

pip install --no-cache <https://developer.download.nvidia.com/compute/redist/jp/v60dp/pytorch/torch-2.2.0a0+81ea7a4.nv24.01-cp310-cp310-linux_aarch64.whl>
# pip install torchvision

See https://github.com/dusty-nv/jetson-containers/tree/master/packages/l4t/l4t-pytorch