Running and Monitoring Training Scripts in Azure Machine Learning

This lesson looks at the process of converting notebooks into scripts and then running those scripts from the command line. This involves converting notebook cells to a script, preparing code for production, and using the Python SDK to create a command job. Jobs’ parameters, metrics, artifacts, and models can be monitored and logged with MLFlow. This lesson looks at the ways this is achieved and how to view the logged elements via scripts and Azure Machine Learning Studio.

Learning Objectives

• Understand how to convert a Jupyter Notebook into a script
• Practice running a script as a command
• Use MLflow to log model training

Intended Audience 

• Students wanting to know how to automate aspects of model training
• Students preparing for the DP-100: Designing and Implementing a Data Science Solution on Azure exam

Prerequisites

• Familiarity with data science concepts
• Some exposure to Python

Running a Script as a Command demo commands

rm -r azure-ml-labs -f
git clone https://github.com/MicrosoftLearning/mslearn-azure-ml.git azure-ml-labs

cd azuer-ml-labs/Labs/08
./setup.sh

pip uninstall azure-ai-ml
pip install azure-ai-ml

git clone https://github.com/MicrosoftLearning/mslearn-azure-ml.git azure-ml-labs