Train, Track and Deploy Regression Model with Databricks, MLFlow and Azure Machine Learning Service
The following 4 Azure Databricks Notebooks are contained in TurbofanML.dbc:
- 01 - NASA_TURBOFAN_PREP - Retrieve NASA Turbofan data, prep and save for additional use
- 02 - NASA_TURBOFAN_ML - Use Spark.ML to predict remaining useful life
- 03 - NASA_TURBOFAN_AUTOML - Employ Azure Machine Learning AutoML to train and register the 'best model' and then deploy to Azure Container Instance
- 04 - NASA_TURBOFAN_MLFLOW - Run a parameterized Notebook to track experiements with Databricks MLFlow, then change input parameter to monitoring experiments with Azure Machine Learning Service by integrating MLFlow with your Azure Machine Learning Workspace.