Skip to content

Latest commit

 

History

History
75 lines (64 loc) · 3.86 KB

install_steps.md

File metadata and controls

75 lines (64 loc) · 3.86 KB

Basic Installation

  1. Etch the Ubuntu Image using the instructions here
  2. Configure the installation (requires a keyboard/mouse) after the reboot.
  3. sudo apt update && sudo apt upgrade
  4. sudo apt install python3-pip python-pip
  5. Reboot Jetson Nano
  6. cd ~
  7. wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-aarch64.sh .
  8. chmod a+x Miniforge3-Linux-aarch64.sh
  9. ./Miniforge3-Linux-aarch64.sh (Do not run as root!)
  10. Log out and log back on
  11. conda config --set auto_activate_base false
  12. sudo apt install python3-h5py libhdf5-serial-dev hdf5-tools libpng-dev libfreetype6-dev
  13. conda create -n jupyter python=3.6
  14. conda activate jupyter 1. conda install matplotlib pandas numpy pillow scipy tqdm scikit-image scikit-learn seaborn cython h5py jupyter ipywidgets -c conda-forge

Optional Installation

General Tips and Tricks

  • Jetson Nano CLI Only:

    • If you're only using the Jetson Nano for non-gui purposes, like as a remote inference/training server. This reduces memory pressure of running your OS, and instead focuses on leaving enough memory for your training/inference operations.
    • jetson_config -p desktop
    • Select [B1] to boot directly to non-gui terminal and require login.
    • NOTE: Only use [B2] if you understand what the consequences of this option.
    • This will trigger a Jetson Nano reboot. Click yes.
  • Other Useful Utilities:

  • Disk Imager:

  • Cleanup:

    • sudo apt autoremove