Preface

Deep Learning (DL) is a relatively new type of machine learning which demonstrates incredible results in tasks such as natural language understanding and computer vision. At times, DL can be more accurate than humans.

Thanks to the proliferation of open source frameworks, publicly available model architectures and pertained models, many people and organizations can successfully apply cutting-edge DL models to their practical use cases. However, developing, training, and deploying DL models also requires highly specialized and costly types of hardware, software stacks, expertise, and management capabilities which may considerably slow down the adoption.

This book focuses on how to engineer and manage Deep Learning workloads on Amazon SageMaker, which allows you to overcome the aforementioned barriers. SageMaker is a sprawling AWS cloud Machine Learning platform with a variety of capabilities. This book does not intend to cover all available SageMaker capabilities in detail, but rather dive deep into the features relevant to DL workloads. We prioritized depth over breadth when writing this book. The goal of this book is to provide you with practical guidelines on how to efficiently implement real-time use cases involving Deep Learning models on Amazon SageMaker.

Since cloud adoption and machine learning adoption are both accelerating, this book may be of interest to a wide audience, from beginners to experienced ML practitioners. Specifically, this book is for ML engineers who work on DL model development and training, and Solutions Architects who are in charge of designing and optimizing DL workloads.

It is assumed that you are familiar with the Python ecosystem, and the principles of Machine Learning and Deep Learning. Familiarity with AWS and practical experience working with it are also helpful.

The complexity of the chapters increases as we move from introductory and overview topics to advanced implementation and optimization techniques. You may skip certain chapters, or select specific topics which are relevant to your specific task at hand.

Most chapters of this book have corresponding code examples so you can develop practical experience working with Amazon SageMaker. It’s recommended that you try to run the code samples yourself, however, you may also review them. We also provide commentary for code samples as part of each chapter.

Please note that running code examples will results in AWS charges. Make sure to check the Amazon SageMaker pricing page for details.

We welcome your feedback and suggestions on this book, and hope that you enjoy your learning journey.