book

TensorFlow 2 Pocket Reference

Name: TensorFlow 2 Pocket Reference
Author: KC Tung
ISBN: 9781492089186

by KC Tung

July 2021

Intermediate to advanced

253 pages

5h 1m

English

O'Reilly Media, Inc.

Read now

Unlock full access

Preface
Conventions Used in This BookUsing Code ExamplesO’Reilly Online LearningHow to Contact UsAcknowledgments
1. Introduction to TensorFlow 2
Improvements in TensorFlow 2Keras APIReusable Models in TensorFlowMaking Commonly Used Operations EasyOpen Source DataWorking with Distributed DatasetsData StreamingData EngineeringTransfer LearningModel StylesMonitoring the Training ProcessDistributed TrainingServing Your TensorFlow ModelImproving the Training ExperienceWrapping Up
2. Data Storage and Ingestion
Streaming Data with Python GeneratorsStreaming File Content with a GeneratorJSON Data StructuresSetting Up a Pattern for FilenamesSplitting a Single CSV File into Multiple CSV FilesCreating a File Pattern Object Using tf.ioCreating a Streaming Dataset ObjectStreaming a CSV DatasetOrganizing Image DataUsing TensorFlow Image GeneratorStreaming Cross-Validation ImagesInspecting Resized ImagesWrapping Up
3. Data Preprocessing
Preparing Tabular Data for TrainingMarking ColumnsEncoding Column Interactions as Possible FeaturesCreating a Cross-Validation DatasetStarting the Model Training ProcessSummaryPreparing Image Data for ProcessingTransforming Images to a Fixed SpecificationTraining the ModelSummaryPreparing Text Data for ProcessingTokenizing TextCreating a Dictionary and Reverse DictionaryWrapping Up
4. Reusable Model Elements
The Basic TensorFlow Hub WorkflowImage Classification by Transfer LearningModel RequirementsData Transformation and Input ProcessingModel Implementation with TensorFlow HubDefining the OutputMapping Output to Plain-Text FormatEvaluation: Creating a Confusion MatrixSummaryUsing the tf.keras.applications Module for Pretrained ModelsModel Implementation with tf.keras.applicationsFine-Tuning Models from tf.keras.applicationsWrapping Up
5. Data Pipelines for Streaming Ingestion
Streaming Text Files with the text_dataset_from_directory FunctionDownloading Text Data and Setting Up DirectoriesCreating the Data PipelineInspecting the DatasetSummaryStreaming Images with a File List Using the flow_from_dataframe MethodDownloading Images and Setting Up DirectoriesCreating the Data Ingestion PipelineInspecting the DatasetBuilding and Training the tf.keras ModelStreaming a NumPy Array with the from_tensor_slices MethodLoading Example Data and LibrariesInspecting the NumPy ArrayBuilding the Input Pipeline for NumPy DataWrapping Up
6. Model Creation Styles
Using the Symbolic APILoading the CIFAR-10 ImagesInspecting Label DistributionInspecting ImagesBuilding a Data PipelineBatching the Dataset for TrainingBuilding the ModelUnderstanding InheritanceUsing the Imperative APIDefining a Model as a ClassChoosing the APIUsing the Built-In Training LoopCreating and Using a Custom Training LoopCreating the Elements of the LoopPutting the Elements Together in a Custom Training LoopWrapping Up
7. Monitoring the Training Process
Callback ObjectsModelCheckpointEarlyStoppingSummaryTensorBoardInvoking TensorBoard by Local Jupyter NotebookInvoking TensorBoard by Local Command TerminalInvoking TensorBoard by Colab NotebookVisualizing Model Overfitting Using TensorBoardVisualizing the Learning Process Using TensorBoardWrapping Up
8. Distributed Training
Data ParallelismAsynchronous Parameter ServerSynchronous AllreduceUsing the Class tf.distribute.MirroredStrategySetting Up Distributed TrainingUsing a GPU Cluster with tf.distribute.MirroredStrategySummaryThe Horovod APICode Pattern for Implementing the Horovod APIEncapsulating the Model ArchitectureEncapsulating the Data Separation and Sharding ProcessesParameter Synchronization Among WorkersModel Checkpoint as a CallbackDistributed Optimizer for Gradient AggregationDistributed Training Using the Horovod APIWrapping Up
9. Serving TensorFlow Models
Model SerializationSaving a Model to h5 FormatSaving a Model to pb FormatSelecting the Model FormatTensorFlow ServingRunning TensorFlow Serving with a Docker ImageWrapping Up

10. Improving the Modeling Experience: Fairness Evaluation and Hyperparameter Tuning
Model FairnessModel Training and ScoringFairness EvaluationRendering Fairness IndicatorsHyperparameter TuningInteger Lists as HyperparametersItem Choice as HyperparametersFloating-Point Values as HyperparametersEnd-to-End Hyperparameter TuningImport Libraries and Load DataWrapping Up
Index

Overview

This easy-to-use reference for TensorFlow 2 design patterns in Python will help you make informed decisions for various use cases. Author KC Tung addresses common topics and tasks in enterprise data science and machine learning practices rather than focusing on TensorFlow itself.

When and why would you feed training data as using NumPy or a streaming dataset? How would you set up cross-validations in the training process? How do you leverage a pretrained model using transfer learning? How do you perform hyperparameter tuning? Pick up this pocket reference and reduce the time you spend searching through options for your TensorFlow use cases.

Understand best practices in TensorFlow model patterns and ML workflows
Use code snippets as templates in building TensorFlow models and workflows
Save development time by integrating prebuilt models in TensorFlow Hub
Make informed design choices about data ingestion, training paradigms, model saving, and inferencing
Address common scenarios such as model design style, data ingestion workflow, model training, and tuning

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

Publisher Resources

ISBN: 9781492089179Errata Page

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills