Hands-On Transfer Learning with TensorFlow 2.x

  • Packt Publishing Limited
  • 2020
  • Paperback
  • 312
  • No language defined
  • Edition not defined
  • 9781838550349
0

Get to grips with the fundamentals of transfer learning and learn how to implement them using real-world examples in TensorFlow 2.0

Key Features

* Apply transfer learning techniques to train efficient deep learning models
* Build transfer learning models from scratch using the power of TensorFlow 2.0
* Overcome the challenges faced while applying transfer learning in computer vision and NLP

Book Description

Transfer learning is about using tried-and-tested models to solve similar problems across the domain, with some customization. This book will be your ultimate guide to implementing transfer learning with TensorFlow 2.0.

Starting with the fundamentals of transfer learning, this book will take you through related topics in machine learning and deep learning. You'll then be able to define your own convolutional neural network models from scratch to solve image classification problems. With the help of sample architectures such as YOLO and RetinaNet, you'll see how object localization works. As you advance, you'll also get to grips with image segmentation architectures such as R-CNN. This book will help you understand the strategy of applying transfer learning in natural language processing (NLP) tasks such as language modeling, sentiment analysis, and language translation. Finally, you'll cover how to apply stochastic gradient descent (SGD) with restart, differential learning rates, and data augmentation techniques to fine-tune models. You'll not only know the different techniques but also understand when to apply them.

By the end of the book, you'll be able to perform core transfer learning techniques across deep learning domains using TensorFlow 2.x.

What you will learn

* Grasp the principle concepts of machine learning and deep learning
* Explore and apply transfer learning in computer vision and NLP problems
* Understand how object localization works using architectures such as YOLO and RetinaNet
* Discover how image segmentation works using R-CNN
* Build a basic Seq2Seq model from scratch
* Understand and apply the data augmentation techniques
* Apply fine-tuning techniques to improve image classification models

Who This Book Is For

scientists, machine learning engineers, AI developers, and researchers who want to build transfer learning models using state-of-the-art methodologies will find this book useful. Familiarity with basic machine learning and exposure to Python programming is all that you need to get started with this book. Knowledge of deep learning and neural networks will be helpful.

333,00 kr.