Train and Deploy a Machine Learning Model with Azure Machine Learning (DP-3007)

Train and Deploy a Machine Learning Model with Azure Machine Learning (DP-3007) Benefits

• Upon successful completion of this course, students will master essential skills to: 

  • Make data available in Azure Machine Learning. 
  • Work with compute targets in Azure Machine Learning. 
  • Run a training script as a command job in Azure Machine Learning. 
  • Track model training with MLflow in jobs. 
  • Register an MLflow model in Azure Machine Learning. 
  • Deploy a model to a managed online endpoint. 

• Training Prerequisites

  To maximize the benefits of this course, participants should have familiarity with the data science process. While the course doesn't delve deeply into data science concepts, a basic understanding is recommended. Additionally, familiarity with Python is essential, as the course focuses on utilizing the Python SDK for interacting with Azure Machine Learning.

Azure Machine Learning DP-3007 training course Outline

Module 1: Make Data Available in Azure Machine Learning 

• Introduction 
• Understand URIs 
• Create a datastore 
• Create a data asset 

Exercise: Make data available in Azure Machine Learning 

Module 2: Work with Compute Targets in Azure Machine Learning 

• Introduction 
• Choose the appropriate compute target 
• Create and use a compute instance 
• Create and use a compute cluster 

Exercise: Work with compute resources 

Module 3: Work with Environments in Azure Machine Learning 

• Introduction 
• Understand environments 
• Explore and use curated environments 
• Create and use custom environments 

Exercise: Work with environments 

Module 4: Run a Training Script as a Command Job in Azure Machine Learning 

• Introduction 
• Convert a notebook to a script 
• Run a script as a command job 
• Use parameters in a command job 

Exercise: Run a training script as a command job 

Module 5: Track Model Training with MLflow in Jobs 

• Introduction 
• Track metrics with MLflow 
• View metrics and evaluate models 

Exercise: Use MLflow to track training jobs 

Module 6: Register an MLflow Model in Azure Machine Learning 

• Introduction 
• Log models with MLflow 
• Understand the MLflow model format 
• Register an MLflow model 

Exercise: Log and register models with MLflow 

Module 7: Deploy a Model to a Managed Online Endpoint 

• Introduction 
• Explore managed online endpoints 
• Deploy your MLflow model to a managed online endpoint 
• Deploy a model to a managed online endpoint 
• Test managed online endpoints 

Exercise: Deploy an MLflow model to an online endpoint