Skip to content

Latest commit

 

History

History

azure_databricks_version

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 

Azure databricks version instructions

  • create a new azure databricks workspace from portal.azure.com
  • launch the workspace
  • create a new cluster
  • ingest data (training data in ../data)
  • import and run the ML.ipynb notebook in a databricks notebook/ workspace
  • (optional) look at the experiment results in the MLflow tab
  • click "use model for inference" from models tab and copy the code from "query endpoint" section
  • create a new token (click on user settings and generate new token) and save it as DATABRICKS_TOKEN in your environment variables
  • run python test_endpoint.py in your local python environment to test the endpoint, if you get a prediction as a response, you are good to go!
  • run streamlit run dashboard.py to launch the dashboard

Clean up

  • delete the cluster and workspace from azure portal