Delve into ML.NET to build and train models for various machine learning tasks. Explore key features, advanced capabilities like deep learning, and integration with TensorFlow.

Using ML.NET to Build Machine Learning Models for AI Solutions.png

mlnet.tar.gz

Command Line

Model Execution

In this course, you will learn how to use ML.NET, which is a tool based on .NET architecture. It consists of a library and command line utility used for building machine learning models. It is so convenient to work with that a developer having little or no background in machine learning and data science can use it to build complex machine learning models.

You will start with an overview of the key ML.NET features and the fundamentals of machine learning. Then, you will go through all types of built-in tasks supported by ML.NET, followed by more advanced ML.NET capabilities, such as deep learning and interoperability with external tools, such as TensorFlow.

By the end of the course, you will be able to use ML.NET to build and train models capable of performing a wide range of machine learning tasks. You will be able to use all the key features of ML.NET and fully integrate it into your apps.

Using ML.NET to Build Machine Learning Models

Which step is automated by AutoML?

AutoML is available by default without having to install any additional packages.

Which NuGet package enables AutoML in ML.NET?

Which NuGet package enables AutoML in ML.NET?

It doesn't. We need to tell it what data types each column represents.

It examines several rows of data and infers the data type for each column.

It assumes that all columns are strings unless we tell it otherwise.

It looks at the first row of data and infers the data types from it.

How does AutoML determine the data types of the features?

How does AutoML determine the data types of the features?

The algorithm that selects the best training algorithm

What is an experiment in the context of AutoML?

What is an experiment in the context of AutoML?

To exclude specific algorithms that we know don't perform particularly well

In which case would you want to apply additional configuration to an AutoML pipeline?

In which case would you want to apply additional configuration to an AutoML pipeline?

Introduction to ML.NET

Machine Learning Fundamentals

Selecting a Problem for Machine Learning

Built-In Supervised Learning Tasks in ML.NET

Built-In Unsupervised Learning Tasks in ML.NET

Deep Learning and Neural Networks

Automating Machine Learning Tasks with AutoML

Saving and Consuming Machine Learning Models

Building an Image Analysis Application

Wrapping Up

Appendix

Quiz Yourself on AutoML