Is this page helpful?

Frequently Asked Questions for Installing AI Workbench#

General#

This section answers frequently asked questions about installing AI Workbench.

Installing on Windows#

What Gets Installed and Where?#

AI Workbench installs files on both the Windows side and into the WSL2 distribution NVIDIA-Workbench.

AI Workbench installs no software dependencies on the Windows side other than files associated with Docker Desktop (optional).

Windows Side

For Windows, the following folders contains files that populate Program Files or AppData:

  • C:Program FilesNVIDIA AI Workbench - The AI Workbench Desktop Application (Desktop App) and related files.

  • C:Users<user-name>AppDataLocal

    • NVIDIA CorporationAI Workbench - This folder contains various configuration files and the credential-manager binary.

    • NVIDIA CorporationAI Workbench Distro - This folder contains the image file for the WSL2 Ubuntu distribution.

    • nvidia-ai-workbench-updater - This folder contains the installer and updater files.

  • C:Users<user-name>AppDataRoamingnvidia-ai-workbench - This folder contains logs, local state file, and a lock file.

NVIDIA-Workbench

The other files are AI Workbench related installation files and include:

  • Software dependencies such as Git, Git-LFS, Podman (optional), and NVIDIA Container Toolkit files if host has a GPU and Podman is selected.

  • Application files and binaries that specifically go into the folder /home/workbench/.nvwb.

What if Docker Desktop is Already Installed?#

If Docker Desktop is already installed, the installer will default to the existing installation and won’t ask you to select a container runtime.

How Do I Change to Podman Instead of Docker Desktop?#

This is a complicated process if you have Docker Desktop installed and want to switch to Podman. At this time, we don’t yet have a simple way to switch the runtimes. If Docker Desktop is installed, AI Workbench uses it as the default container runtime and AI Workbench won’t let you switch to Podman, unless you uninstall Docker Desktop first.

To use Podman instead of Docker Desktop:

  1. Completely close and quit all instances of AI Workbench.

  2. Uninstall Docker Desktop from your Windows system

  3. Go into the NVIDIA AI Workbench distro

  4. Open the file /home/workbench/.nvwb/config.yaml in a text editor

  5. Change the container runtime and build-time files to podman instead of docker

  6. Save the file(s).

  7. Restart AI Workbench and select the Podman container runtime when prompted

Visual Studio Code (VSCode) needs consideration if it is configured to use Docker Desktop and you want to use Podman instead. In this case VS Code settings will also need to be manually updated. For more information, see Using Podman instead of Docker.

Do I Need to Install the NVIDIA GPU Drivers even if my Windows System Has an NVIDIA GPU?#

Yes. You must install the NVIDIA GPU drivers on your own. We highly recommend using the NVIDIA App to install and manage your NVIDIA GPU drivers.

Why do I need to install a dedicated WSL distribution?#

AI Workbench uses containers, which must run in a WSL distribution, therefore a dedicated distribution is required to avoid interfering with other WSL distributions you may have installed.

Installing on MacOS#

What Gets Installed and Where on a Mac?#

AI Workbench installs the following components on your MacOS:

  • The application binaries: nvwb-cli, wb-svc, and credential-manager into $HOME/.nvwb/bin.

  • Git and Git-LFS: Installed or updated system wide.

  • Docker Desktop or Podman: Installed or updated system wide.

  • Homebrew: Installed or updated system wide.

Can I Install the Workbench CLI Instead of the Desktop App?#

Yes. On macOS you do not need to install the AI Workbench Desktop App.

To install AI Workbench CLI, follow these steps:

  1. Open a terminal, copy, paste and run the following command:

    mkdir -p $HOME/.nvwb/bin && \
curl -L https://workbench.download.nvidia.com/stable/workbench-cli/$(curl -L -s https://workbench.download.nvidia.com/stable/workbench-cli/LATEST)/nvwb-cli-$(uname)-$(uname -m) --output $HOME/.nvwb/bin/nvwb-cli && \
chmod +x $HOME/.nvwb/bin/nvwb-cli && \
sudo -E $HOME/.nvwb/bin/nvwb-cli install

  2. Follow the text-based interface prompts to complete the installation:

    • Use arrow keys to navigate and Enter to select options.

    • Select either Docker or Podman as your container runtime.

Note

You can update the CLI by running the same command as above.

You can uninstall the CLI using the following command:

sudo -E $HOME/.nvwb/bin/nvwb-cli uninstall

Installing on Ubuntu Desktop#

What gets installed and where?#

AI Workbench installs the following components on your system:

  • The application binaries: nvwb-cli, wb-svc, and credential-manager into $HOME/.nvwb/bin.

  • Git and Git-LFS: Installed or updated system wide.

  • Docker or Podman: Installed or updated system wide.

  • (If GPU): NVIDIA GPU Drivers, if not already installed

  • (If GPU): NVIDIA Container Toolkit, installed or updated system wide.

Can I just install the Workbench CLI instead of the Desktop App?#

Yes. On Ubuntu Desktop you do not need to install the Desktop App.

To install AI Workbench CLI, follow these steps:

  1. Open a terminal, copy, paste and run the following command:

    mkdir -p $HOME/.nvwb/bin && \
curl -L https://workbench.download.nvidia.com/stable/workbench-cli/$(curl -L -s https://workbench.download.nvidia.com/stable/workbench-cli/LATEST)/nvwb-cli-$(uname)-$(uname -m) --output $HOME/.nvwb/bin/nvwb-cli && \
chmod +x $HOME/.nvwb/bin/nvwb-cli && \
sudo -E $HOME/.nvwb/bin/nvwb-cli install

  2. Follow the text-based interface prompts to complete the installation:

    • Use arrow keys to navigate and Enter to select options

    • If you have an NVIDIA GPU without drivers, you can choose to install them

    • Select either Docker or Podman as your container runtime

  3. Reboot if prompted (required if GPU drivers were installed)

Note

You can update the CLI by running the same command as above.

You can uninstall the CLI using the following command:

sudo -E $HOME/.nvwb/bin/nvwb-cli uninstall

What if I have the AppImage installed but I want to switch to the Debian package?#

To switch to the Debian package, uninstall (see here) the AppImage and then install the Debian package.

Your projects and locations will be preserved.

How Do I install and use the AppImage?#

Note

The AppImage is not the recommended installation method and will be deprecated in the future.

To install the AppImage, follow these steps:

  1. Install fuse3 and libfuse2 with the command:

    sudo apt-get install fuse3 libfuse2

  2. Download the Ubuntu Desktop App: AI Workbench

  3. Make the AppImage executable with the command:

    chmod +x <local-path-to-AppImage>

  4. Run the AppImage by double-clicking it or running it in a terminal, and then following the prompts. :

  5. Reboot if prompted (required if GPU drivers were installed)

How do I uninstall the AppImage?#

You can uninstall the AppImage by:

  • Running the CLI uninstall command:

    sudo -E $HOME/.nvwb/bin/nvwb-cli uninstall

  • Deleting the downloaded AppImage file manually using the following command:

    rm -f /path/to/NVIDIA-AI-Workbench.AppImage

Installing on Ubuntu Remote#

What Gets Installed/Updated and Where?#

AI Workbench installs the following components on the remote system:

  • The application binaries: nvwb-cli, wb-svc, and credential-manager into $HOME/.nvwb/bin.

  • Git: Installed or updated system wide

  • Docker or Podman: Installed or updated system wide

  • (If GPU): NVIDIA GPU Drivers, if not already installed

  • (If GPU): NVIDIA Container Toolkit, installed or updated system wide

Do I install the Desktop App on a remote?#

No. The AI Workbench Desktop App is not supported on a remote system.

  • You cannot install on a remote with either the Debian package or the AppImage.

  • You must use the CLI for the remote install.

How do I do a non-interactive install using the CLI?#

To do a non-interactive install using the CLI, see Remote Ubuntu Install.