Install AI Workbench Locally on Ubuntu 22.04

Use this documentation to install NVIDIA AI Workbench on your local computer running Ubuntu 22.04. For an overview, see Install AI Workbench.

• Ubuntu 22.04

• The fuse package to support AppImage

• 500 MB of disk space for the AI Workbench application

• 30-40 GB of disk space for containers

• 16 GB of RAM

1. Verify that fuse is installed by running dpkg -s fuse libfuse2. If not, install it by running sudo apt update followed by sudo apt install fuse.

2. Download AI Workbench, make it executable, and then run it. Make it executable in a terminal by running chmod +x local/path/to/download.

   Warning

   Do not run the AI Workbench installation as the root user.

3. (optional) If your computer has an NVIDIA GPU but no drivers installed, AI Workbench detects this and offers to install the latest drivers for you. For more information, see NVIDIA GPU and Drivers.

   • If you select this option, you need to reboot the computer after the drivers are installed.

   • If you want to install the latest drivers yourself, then stop the installation, install them, and reboot the computer.

   • After you reboot, restart the AI Workbench installation by double-clicking the NVIDIA AI Workbench icon on your desktop.

4. Choose Docker or Podman as your container runtime. For more information, see Container runtime options. AI Workbench installs the option you select.

5. (Optional) Connect AI Workbench to your GitHub or GitLab account. If you skip this step now, you must enter Git Author information when you are prompted. If you skip this step now, you can connect to Git later through the Settings menu. For more information, see Connect to Git.

6. AI Workbench opens.

• Do a Quickstart

  • Basic Quickstart

  • Basic CLI Quickstart

  • Hybrid RAG Quickstart

• Explore the NVIDIA Developer Forum

• Learn the Concepts

  • Workbench Projects

  • Overview of Locations

  • Overview of Integrations

  • Applications