Install AI Workbench

Use the documentation in this section to install NVIDIA AI Workbench on your computer. You can install AI Workbench locally, or on a remote computer that you have SSH access to.

AI Workbench is supported on the following operating systems:

When you install AI Workbench, Git, Git Large File Storage (LFS), and other container-related dependencies are installed.

To get help when you install AI Workbench, see the NVIDIA Developer Forum.

Container runtime options

When you install AI Workbench you need to choose a container runtime. Currently, Docker and Podman are supported. Docker is generally considered a better user experience on Windows and macOS, but you might need an organizational license to use Docker Desktop. Podman is free and open source, but comes with the risks associated with open-source software.

NVIDIA GPU and Drivers

You can install AI Workbench on a computer with no GPU, and some tasks never require a GPU. Many AI and data science workloads eventually need GPU acceleration. When you install AI Workbench on a Linux or macOS computer with GPU support, AI Workbench can install the drivers for you. When you install AI Workbench on a Windows computer with GPU support, you must install the NVIDIA drivers manually. If your Windows computer doesn’t already have drivers installed, we recommend the following drivers:

Connect to Git

Each AI Workbench project is stored as a git repository. When you install AI Workbench you are prompted to connect to a Git repository manager. Currently, GitHub and GitLab are supported. You can skip this step during installation and come back to it later, and you can change your Git integration at any time.

To connect to Git

  1. Start AI Workbench by double clicking the icon on your desktop, or by searching for the program and pressing enter. The icon has the NVIDIA logo in it.

  2. After AI Workbench opens, right-click on the AI Workbench icon on the taskbar and click Settings. Alternatively, you can right-click the gray AI Workbench logo in the computer’s system tray (AppIndicator on Ubuntu 22.04, Notification Area on macOS) and click Settings.

    The Workbench Settings window appears.

  3. In the Workbench Settings window, select the Integration page.

  4. Connect (or disconnect) to GitHub or GitLab. You can connect by logging in or by using a personal access token (PAT).

To install AI Workbench, choose the following instructions that correspond to your operating system.

Install AI Workbench

Previous NVIDIA AI Workbench Shared Security Model
Next Install AI Workbench on Windows
© Copyright © 2024, NVIDIA Corporation. Last updated on Apr 29, 2024.