A no-nonsense demo of MLflow on Databricks

1. In Databricks, you can either:
   • Download and import this file as a .py file
   • If using Databricks Repos, click "Create Repo" and pass in the public URL for this Github Repo
2. Create a Cluster with (at least) the following settings:
   • ML runtime (tested on DBR 7.6 ML) Docs
   • Enable Cluster Spark config: spark.databricks.mlflow.autologging.enabled true Docs
3. Run the cells!
   • You will see runs for each model appear in the sidebar on the right.
   • Click one of these to explore the MLflow UI
   • From the Artifacts of a run, click Register Model under a new name (e.g. mlflow101_model)
   • In Model Registry, promote this model to Production, then run the last cell in the notebook using the name of that model (e.g. mlflow101_model)