TensorFlow 2.0 Quick Start Guide电子书

Dedication

About Packt

Why subscribe?

Packt.com

Contributors

About the author

About the reviewers

Packt is searching for authors like you

Preface

Who this book is for

What this book covers

To get the most out of this book

Download the example code files

Download the color images

Conventions used

Get in touch

Reviews

Section 1: Introduction to TensorFlow 2.00 Alpha

Introducing TensorFlow 2

Looking at the modern TensorFlow ecosystem

Installing TensorFlow

Housekeeping and eager operations

Importing TensorFlow

Coding style convention for TensorFlow

Using eager execution

Declaring eager variables

Declaring TensorFlow constants

Shaping a tensor

Ranking (dimensions) of a tensor

Specifying an element of a tensor

Casting a tensor to a NumPy/Python variable

Finding the size (number of elements) of a tensor

Finding the datatype of a tensor

Specifying element-wise primitive tensor operations

Broadcasting

Transposing TensorFlow and matrix multiplication

Casting a tensor to another (tensor) datatype

Declaring ragged tensors

Providing useful TensorFlow operations

Finding the squared difference between two tensors

Finding a mean

Finding the mean across all axes

Finding the mean across columns

Finding the mean across rows

Generating tensors filled with random values

Using tf.random.normal()

Using tf.random.uniform()

Using a practical example of random values

Finding the indices of the largest and smallest element

Saving and restoring tensor values using a checkpoint

Using tf.function

Summary

Keras, a High-Level API for TensorFlow 2

The adoption and advantages of Keras

The features of Keras

The default Keras configuration file

The Keras backend

Keras data types

Keras models

The Keras Sequential model

The first way to create a Sequential model

The second way to create a Sequential model

The Keras functional API

Subclassing the Keras Model class

Using data pipelines

Saving and loading Keras models

Keras datasets

Summary

ANN Technologies Using TensorFlow 2

Presenting data to an ANN

Using NumPy arrays with datasets

Using comma-separated value (CSV) files with datasets

CSV example 1

CSV example 2

CSV example 3

TFRecords

TFRecord example 1

TFRecord example 2

One-hot encoding

OHE example 1

OHE example 2

Layers

Dense (fully connected) layer

Convolutional layer

Max pooling layer

Batch normalization layer and dropout layer

Softmax layer

Activation functions

Creating the model

Gradient calculations for gradient descent algorithms

Loss functions

Summary

Section 2: Supervised and Unsupervised Learning in TensorFlow 2.00 Alpha

Supervised Machine Learning Using TensorFlow 2

Supervised learning

Linear regression

Our first linear regression example

The Boston housing dataset

Logistic regression (classification)

k-Nearest Neighbors (KNN)

Summary

Unsupervised Learning Using TensorFlow 2

Autoencoders

A simple autoencoder

Preprocessing the data

Training

Displaying the results

An autoencoder application – denoising

Setup

Preprocessing the data

The noisy images

Creating the encoding layers

Creating the decoding layers

Model summary

Model instantiation, compiling, and training

Denoised images

TensorBoard output

Summary

Section 3: Neural Network Applications of TensorFlow 2.00 Alpha

Recognizing Images with TensorFlow 2

Quick Draw – image classification using TensorFlow

Acquiring the data

Setting up our environment

Preprocessing the data

Creating the model

Training and testing the model

TensorBoard callback

Saving, loading, and retesting the model

Saving and loading NumPy image data using the .h5 format

Loading and inference with a pre-trained model

CIFAR 10 image classification using TensorFlow

Introduction

The application

Summary

Neural Style Transfer Using TensorFlow 2

Setting up the imports

Preprocessing the images

Viewing the original images

Using the VGG19 architecture

Creating the model

Calculating the losses

Performing the style transfer

Final displays

Summary

Recurrent Neural Networks Using TensorFlow 2

Neural network processing modes

Recurrent architectures

An application of RNNs

The code for our RNN example

Building and instantiating our model

Using our model to get predictions

Summary

TensorFlow Estimators and TensorFlow Hub

TensorFlow Estimators

The code

TensorFlow Hub

IMDb (database of movie reviews)

The dataset

The code

Summary

Converting from tf1.12 to tf2

Other Books You May Enjoy

Leave a review - let other readers know what you think