Create and Publish Pipelines for Batch Inferencing with Azure
This course will teach you how to use the Azure Machine Learning service to build and run ML pipelines using the drag-and-drop designer interface. You will cover the publishing and deployment of pipelines for batch and real-time inferences.
What you'll learn
A machine learning model goes through a number of stages in its lifecycle; from training, to evaluation, through deployment and then maintenance. While there are a number of tools available for these stages, their management can become overwhelming even for the seasoned ML engineer.
In this course, Create and Publish Pipelines for Batch Inferencing with Azure, you'll experience an intuitive and easy-to-maintain environment for all things ML and focus on building and running pipelines for batch inferences:
- discover the Azure ML service and the breadth of features it has to offer when it comes to building and managing ML models
- explore a number of data transformations which can be applied to a dataset by simply dragging and dropping various modules into the pipeline
- see that the handling of missing values, the standardization of numeric features as well as one-hot encoding for categorical fields can all be accomplished without writing a line of code
- use the pipeline to make predictions on new data
Table of contents
- Course Prerequisites and Outline 3m
- Introducing Azure Machine Learning 7m
- Datasets in Azure ML 3m
- Azure ML Terms and Concepts 7m
- Demo: Creating an Azure ML Workspace 7m
- Demo: Create a Compute Cluster 3m
- Demo: Exploring the Designer 4m
- Demo: Loading and Configuring a Dataset 9m
- Demo: Summarizing a Dataset 5m
- Demo: Running a Pipeline 3m
- Common Machine Learning Workflows 8m
- Demo: Marking Columns as Categorical 5m
- Demo: Handling Missing Data 8m
- Demo: Applying One-hot Encoding 4m
- Demo: Standardizing Numeric Fields 5m
- Demo: Creating Training and Test Sets 5m
- Demo: Training and Evaluating a Model 5m
- Demo: Examining the Evaluation Metrics 7m
- Hyperparameter Tuning 4m
- Demo: Implementing Hyperparameter Tuning 6m
- Demo: Evaluating Combinations of Hyperparameters 6m
- Batch Inference Pipelines 3m
- Demo: Re-building a Model Training Pipeline 4m
- Demo: Creating and Analyzing a Batch Inference Pipeline 7m
- Demo: Publishing and Running a Batch Inference Pipeline 8m
In this Azure tutorial, you will get an overview of the Azure ML service, learn about a number of data transformations, and how to use the pipeline to make predictions.
Azure Machine Learning is a cloud-based environment you can use to train, deploy, automate, manage, and track ML models.
This course is for anyone who wants to learn Azure ML and create and publish their own pipelines for batch and real-time inferences.
A machine learning model is a file that has been trained to recognize certain types of patterns. You train a model over a set of data, providing it an algorithm that it can use to reason over and learn from certain data.
Data pipelines are sets of data processing elements connected in a series where the output of one element is the input of the next one.