Ashish Ranjan Jha - Mastering PyTorch

Build powerful neural network architectures using advanced PyTorch 1.x features

Ashish Ranjan Jha

BIRMINGHAMMUMBAI

Copyright 2021 Packt Publishing

All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews.

Every effort has been made in the preparation of this book to ensure the accuracy of the information presented. However, the information contained in this book is sold without warranty, either express or implied. Neither the author, nor Packt Publishing or its dealers and distributors, will be held liable for any damages caused or alleged to have been caused directly or indirectly by this book.

Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.

Group Product Manager: Kunal Parikh

Publishing Product Manager: Devika Battike

Senior Editor: Roshan Kumar

Content Development Editor: Tazeen Shaikh

Technical Editor: Sonam Pandey

Copy Editor: Safis Editing

Project Coordinator: Aishwarya Mohan

Proofreader: Safis Editing

Indexer: Priyanka Dhadke

Production Designer: Nilesh Mohite

First published: February 2021

Production reference: 1140121

Published by Packt Publishing Ltd.

Livery Place

35 Livery Street

Birmingham

B3 2PB, UK.

ISBN 978-1-78961-438-1

www.packt.com

To my mother and best-friend - Rani Jha, my father and idol - Bibhuti Bhushan Jha, for their sacrifices and constant support and for being the driving forces of my life and career. Without their love, none of this would matter. To my sisters, Sushmita, Nivedita, and Shalini, for teaching me what and what not to do in life.

Packt.com

Subscribe to our online digital library for full access to over 7,000 books and videos, as well as industry leading tools to help you plan your personal development and advance your career. For more information, please visit our website.

• Spend less time learning and more time coding with practical eBooks and videos from over 4,000 industry professionals
• Improve your learning with Skill Plans built especially for you
• Get a free eBook or video every month
• Fully searchable for easy access to vital information
• Copy and paste, print, and bookmark content

Did you know that Packt offers eBook versions of every book published, with PDF and ePub files available? You can upgrade to the eBook version at for more details.

At www.packt.com, you can also read a collection of free technical articles, sign up for a range of free newsletters, and receive exclusive discounts and offers on Packt books and eBooks.

I am happy to know that Ashish, who was my student on the artificial neural networks course 8 years ago at IIT Roorkee, has now authored this hands-on book that covers a range of deep learning topics in reasonable depth.

Learning by coding is something every deep learning enthusiast wants to undertake, but tends to leave half way through. The effort needed to go through documentation and extract useful information to run deep learning projects is cumbersome. I have seen far too many students become frustrated during the process. There are tons of resources available for any beginner to become an expert. However, it is easy for any beginner to lose sight of the learning task while trying to strike a balance between concept-oriented courses and the coding-savvy approach of many academic programs.

PyTorch is uniquely placed as being pythonic and very flexible. It is appealing both to beginners who have just started coding machine learning models and to experts who like to meddle in the finer parameters of model designing and training. PyTorch is one library I am happy to recommend to any enthusiast, regardless of their level of expertise.

The best way to learn machine learning and deep learning models is by practicing coding in PyTorch. This book navigates the world of deep learning through PyTorch in a very engaging way. It starts from the basic building blocks of deep learning. The visual appeal of learning the data pipeline is one of its strong points. The PyTorch modules used for model building and training are introduced in the simplest of ways. Any student will appreciate the hands-on approach of this book. Every concept is explained through codes, and every step of the code is well documented. It should not be assumed that this book is just for beginners. Instead, any beginner can become an expert by following this book.

Starting from basic model building, such as the popular VGG16 or ResNet, to advanced topics, such as AutoML and distributed learning, all these aspects are covered here. The book further encompasses concepts such as AI explainability, deep reinforcement learning, and GANs. The exercises in this book range from building an image captioning model to music generation and neural style transfer models, as well as building PyTorch model servers in production systems. This helps you to prepare for any niche deep learning ventures.

I recommend this book to anyone who wants to master PyTorch for deploying deep learning models with the latest libraries.

Dr. Gopinath Pillai Head Of Department, Electrical Engineering, IIT Roorkee

Ashish Ranjan Jha received his bachelor's degree in electrical engineering from IIT Roorkee (India), his master's degree in computer science from EPFL (Switzerland), and an MBA degree from the Quantic School of Business (Washington). He received distinctions in all of his degrees. He has worked for a variety of tech companies, including Oracle and Sony, and tech start-ups, such as Revolut, as a machine learning engineer.

Aside from his years of work experience, Ashish is a freelance ML consultant, an author, and a blogger (datashines). He has worked on products/projects ranging from using sensor data for predicting vehicle types to detecting fraud in insurance claims. In his spare time, Ashish works on open source ML projects and is active on StackOverflow and kaggle (arj7192).

Javier Abascal Carrasco has a master's degree in telecommunication engineering from the University of Seville (Spain). He also studied abroad at TU Dresden (Germany) and Thomas College (ME, USA), where he obtained his MBA. Since his career started, Javier has been passionate about the world of data and analytics. He has had the chance to work with and help all manner of companies, ranging from small start-ups to big corporations, including the consulting firm EY and Facebook. In addition, for the last 3 years, he has been a part-time lecturer on the data science space. He truly believes that PyTorch is bringing a new, fresh style to programming and work involving deep learning, generating a friendly competitor landscape in relation to TensorFlow.