Windows Installation

Note

Installing the Desktop Application will also automatically install the Command Line UI onto your system.

Before starting you should have a GitHub.com or a GitLab.com account.

You should also verify that your version of Windows supports WSL2. Windows 11 and Windows 10 builds higher than 19041 support WSL2.

  1. Go to the EA page, and check the box to agree to the EULA.

  2. In the Application Files section, select the installer for Windows and download it.

    1. Download and double-click the file.

    2. Authorize Windows to run the application.

    3. An application window will open to guide you through the installation.

    4. At various points, Windows will request that you verify the changes to your system during the installation. Say yes to each of them.

  3. Click the Begin Installation Button

    image201.png


  4. An NVIDIA AI Workbench icon will appear on your desktop.

Important

If WSL2 is not yet installed on your system, AI Workbench will install it for you. Click Next at the WSL2 Required screen to enable a virtualization feature required for WSL2. You will then need to reboot.

image21.png

When your system starts again, AI Workbench will install WSL2. This automatically installs a default Ubuntu distro called Ubuntu. This distro will not be used by AI Workbench. Close the terminal if one opens.

Instead, AI Workbench will install another Ubuntu 22.04 distro, make some changes to it and name it NVIDIA-Workbench. This is where the dependencies and most of the application files will be installed.

  1. AI Workbench requires a container runtime. Select either Podman or Docker and click Next to continue.

    Warning

    Some considerations when selecting Podman vs Docker on Windows:

    • There is a known issue with Docker Desktop on the Windows Installation (see Known Issues). Podman is recommended for Windows while this issue is being investigated.

    • Podman is open source whereas Docker is not. You may need an organizational license to use Docker Desktop, depending on your organization. However, Podman is free to use.

    image221.png

  2. The selected container runtime and the other dependencies will be installed and you can follow the progress in the application window.

  3. When all dependencies are installed the Next button will turn green. Click it to continue.

    image231.png


  4. Select either GitHub or GitLab and click Next. You can also skip this part and connect to a Git server later.

    • GitHub

      An authentication code will appear in the application window and a browser should open to a page where you can enter the code.

      Click the Authorize NVIDIA button to allow AI Workbench to access your namespace.

      image261.png

    • GitLab

      A URL will appear in the application window and a browser should open to a page where you can authorize AI Workbench to access your GitLab repositories.

      image271.png

Next Steps:
Previous Bring Your Own Container (BYOC)
Next macOS Installation
© Copyright 2023-2024, NVIDIA. Last updated on Jan 21, 2024.