Skip to content

Latest commit

 

History

History
38 lines (30 loc) · 1.79 KB

Setup.md

File metadata and controls

38 lines (30 loc) · 1.79 KB
Raw

Setup for Responsible AI

The goal of this document is to get your Responsible AI components registered and ready for use in your Azure Machine Learning Workspace

Note: The APIs for the latest AzureML SDK are still in preview and subject to change.

Prerequisites

In the following, we assume that you have access to an AzureML workspace.

Setup through Cloud shell

  1. Go to https://shell.azure.com
  2. Choose a bash shell
  3. (first time) Select a subscription to create a storage account and Microsoft Azure Files share.
  4. (first time) Select "Create storage"
  5. The cloud shell should have AML CLI pre-installed. To ensure you have the latest cmdlets installed run az extension add -n ml -y
  6. Make sure your setup is working with any of the list commands: az -h
  7. Install MiniConda
  8. Restart your shell, and make sure that the conda commands are available
  9. Clone the git repo
git clone https://github.com/Azure/RAI-vNext-Preview
cd RAI-vNext-Preview
  1. Run the install script
./quick-setup.bash <CONDA-ENV-NAME> <SUBSCRIPTION-ID> <RESOURCEGROUP-NAME> <WORKSPACE-NAME>
  1. Validate that the setup worked by checking that the components are showing up under your "custom components" folder in ml.azure.com image If your workspace does not have a Components tab, go to the Pipelines tab, and create a new pipeline. In the pipeline designer, you should be able to find the RAI components.

FAQ

Cloud Shell

If you accidentally minimize your cloud shell, you can retrieve it by clicking here image