Skip to content

Latest commit

 

History

History
34 lines (27 loc) · 1.89 KB

README.md

File metadata and controls

34 lines (27 loc) · 1.89 KB
Raw

MLflow UI on AzureML

Deploy To Azure Visualize

Run MLflow UI on AzureML Compute Instances as a custom application

Instructions

ARM Template

  1. Deploy the above ARM template
    1. Select the existing Workspace name
    2. Select a name for the newly created Compute Instance that will host the UI
  2. Navigate to the newly created CI's terminal login
  3. Run az login and follow the instructions to log in
  4. That's it! Navigate to the mlflow ui app created on the CI in your browser! ui

Manual

  1. Compute Instance in the Workspace of your choice
  2. Custom Application in the above CI that references this repo's docker image
    1. Set docker image to ghcr.io/akshaya-a/azureml-mlflow-ui:main
    2. Add /home/azureuser/.azure : /home/azureuser/.azure as a Bind Mount
    3. Set MLFLOW_TRACKING_URI to the AML Workspace's tracking uri (copy from the Azure Portal)
    4. Set HOME to /home/azureuser
    5. Expose 5001 on both Target + Published ports
  3. Navigate to the newly created CI's terminal login
  4. Run az login and follow the instructions to log in
  5. That's it! Navigate to the mlflow ui app created on the CI in your browser! ui