Linked e-resources
Details
Table of Contents
Intro
Table of Contents
About the Author
About the Technical Reviewer
Acknowledgments
Preface
Chapter 1: TensorFlow Jump Start
What Is TensorFlow 2.0?
TensorFlow 2.x Platform
Training
Data Preparation
Designing Model
Distribution Strategy
Analysis
Model Saving
Deployment
What TensorFlow 2.x Offers?
The tf.keras in TensorFlow
Eager Execution
Distribution
TensorBoard
Vision Kit
Voice Kit
Edge TPU
Pre-trained Models for AIY Kits
Data Pipelines
Installation
Installation
Docker Installation
No Installation
Testing
Summary
Chapter 2: A Closer Look at TensorFlow
A Trivial Machine Learning Application
Creating Colab Notebook
Imports
Importing TensorFlow 2.x
Importing numpy
Setting Up Data
Defining Neural Network
Compiling Model
Training Network
Examining Training Output
Predicting
Full Source Code
Binary Classification in TensorFlow
Setting Up Project
Imports
Mounting Google Drive
Loading Data
Shuffling Data
Examining Data
Data Preprocessing
Checking Nulls
Selecting Features and Labels
Encoding Categorical Columns
Scaling Numerical Values
Creating Training and Testing Datasets
Defining ANN
Compiling Model
Model Training
Performance Evaluation
Predicting on Test Data
Confusion Matrix
Predicting on Unseen Data
Full Source Code
Summary
Chapter 3: Deep Dive in tf.keras
Getting Started
Functional API for Model Building
Sequential Models
Model Subclassing
Predefined Layers
Custom Layers
Saving Models
Whole-Model Saving
Export to SavedModel Format
Saving Architecture
Saving Weights
Saving to JSON
Convolutional Neural Networks
Image Classification with CNN
Creating Project
Image Dataset
Loading Dataset
Creating Training/Testing Datasets
Preparing Data for Model Training
Creating Validation Dataset
Augmenting Data
Model Development
Train/Evaluate/Display Function
Predict Function
Defining Models
A Model with 2 Convolutional Layers
Model_2 with 4 Convolutional Layers
Third Model: 6 Convolutional layers with 32, 64 and 128 filters respectively
Fourth Model: Addition of dropout layer
Model 5
Saving Model
Predicting Unseen Images
Summary
Chapter 4: Transfer Learning
Knowledge Transfer
TensorFlow Hub
Pre-trained Modules
Using Modules
ImageNet Classifier
Setting Up Project
Classifier URL
Creating Model
Preparing Images
Loading Label Mappings
Displaying Prediction
Listing All Classes
Result Discussions
Dog Breed Classifier
Project Description
Creating Project
Loading Data
Setting Up Images and Labels
Preprocessing Images
Processing Image
Associating Labels to Images
Creating Data Batches
Display Function for Images
Selecting Pre-trained Model
Defining Model
Table of Contents
About the Author
About the Technical Reviewer
Acknowledgments
Preface
Chapter 1: TensorFlow Jump Start
What Is TensorFlow 2.0?
TensorFlow 2.x Platform
Training
Data Preparation
Designing Model
Distribution Strategy
Analysis
Model Saving
Deployment
What TensorFlow 2.x Offers?
The tf.keras in TensorFlow
Eager Execution
Distribution
TensorBoard
Vision Kit
Voice Kit
Edge TPU
Pre-trained Models for AIY Kits
Data Pipelines
Installation
Installation
Docker Installation
No Installation
Testing
Summary
Chapter 2: A Closer Look at TensorFlow
A Trivial Machine Learning Application
Creating Colab Notebook
Imports
Importing TensorFlow 2.x
Importing numpy
Setting Up Data
Defining Neural Network
Compiling Model
Training Network
Examining Training Output
Predicting
Full Source Code
Binary Classification in TensorFlow
Setting Up Project
Imports
Mounting Google Drive
Loading Data
Shuffling Data
Examining Data
Data Preprocessing
Checking Nulls
Selecting Features and Labels
Encoding Categorical Columns
Scaling Numerical Values
Creating Training and Testing Datasets
Defining ANN
Compiling Model
Model Training
Performance Evaluation
Predicting on Test Data
Confusion Matrix
Predicting on Unseen Data
Full Source Code
Summary
Chapter 3: Deep Dive in tf.keras
Getting Started
Functional API for Model Building
Sequential Models
Model Subclassing
Predefined Layers
Custom Layers
Saving Models
Whole-Model Saving
Export to SavedModel Format
Saving Architecture
Saving Weights
Saving to JSON
Convolutional Neural Networks
Image Classification with CNN
Creating Project
Image Dataset
Loading Dataset
Creating Training/Testing Datasets
Preparing Data for Model Training
Creating Validation Dataset
Augmenting Data
Model Development
Train/Evaluate/Display Function
Predict Function
Defining Models
A Model with 2 Convolutional Layers
Model_2 with 4 Convolutional Layers
Third Model: 6 Convolutional layers with 32, 64 and 128 filters respectively
Fourth Model: Addition of dropout layer
Model 5
Saving Model
Predicting Unseen Images
Summary
Chapter 4: Transfer Learning
Knowledge Transfer
TensorFlow Hub
Pre-trained Modules
Using Modules
ImageNet Classifier
Setting Up Project
Classifier URL
Creating Model
Preparing Images
Loading Label Mappings
Displaying Prediction
Listing All Classes
Result Discussions
Dog Breed Classifier
Project Description
Creating Project
Loading Data
Setting Up Images and Labels
Preprocessing Images
Processing Image
Associating Labels to Images
Creating Data Batches
Display Function for Images
Selecting Pre-trained Model
Defining Model