book

The Applied TensorFlow and Keras Workshop

by Harveen Singh Chadha, Luis Capelo, Abhranshu Bagchi, Achint Chaudhary, Vishal Chauhan, Alexis Rutherford, Subhash Sundaravadivelu

July 2020

Intermediate to advanced

174 pages

3h 1m

English

Packt Publishing

Read now

Unlock full access

About the BookAudienceAbout the ChaptersConventionsCode PresentationSetting up Your EnvironmentInstallationInstalling Python on WindowsInstalling Python on LinuxInstalling Python on macOSInstalling pipJupyter NotebookJupyterLabInstalling LibrariesAccessing the Code Files
IntroductionWhat are Neural Networks?Successful Applications of Neural NetworksWhy Do Neural Networks Work So Well?Representation LearningFunction ApproximationLimitations of Deep LearningInherent Bias and Ethical ConsiderationsCommon Components and Operations of Neural NetworksConfiguring a Deep Learning EnvironmentSoftware Components for Deep LearningPython 3TensorFlowKerasTensorBoardJupyter Notebook, Pandas, and NumPyExercise 1.01: Verifying the Software ComponentsExploring a Trained Neural NetworkThe MNIST DatasetTraining a Neural Network with TensorFlowExercise 1.02: Training a Neural Network Using the MNIST DatasetTesting Network Performance with Unseen DataActivity 1.01: Training a Neural Network with Different HyperparametersSummary
IntroductionChoosing the Right Model ArchitectureConvolutional Neural Networks (CNNs)Recurrent Neural Networks (RNNs)Long Short-Term Memory (LSTM) NetworksGenerative Adversarial NetworksDeep Reinforcement Learning (DRL)Data NormalizationZ-ScorePoint-Relative NormalizationMaximum and Minimum NormalizationStructuring Your ProblemJupyter NotebookExercise 2.01: Exploring the Bitcoin Dataset and Preparing Data for a ModelUsing Keras as a TensorFlow InterfaceModel ComponentsExercise 2.02: Creating a TensorFlow Model Using KerasFrom Data Preparation to ModelingTraining a Neural NetworkReshaping Time Series DataMaking PredictionsOverfittingActivity 2.01: Assembling a Deep Learning SystemSummary
IntroductionProblem CategoriesLoss Functions, Accuracy, and Error RatesDifferent Loss Functions, Same ArchitectureUsing TensorBoardImplementing Model Evaluation MetricsEvaluating the Bitcoin ModelOverfittingModel PredictionsInterpreting PredictionsExercise 3.01: Creating an Active Training EnvironmentHyperparameter OptimizationLayers and Nodes – Adding More LayersAdding More NodesLayers and Nodes – ImplementationEpochsEpochs – ImplementationActivation FunctionsLinear (Identity) FunctionsHyperbolic Tangent (Tanh) FunctionRectified Linear Unit FunctionsActivation Functions – ImplementationRegularization StrategiesL2 RegularizationDropoutRegularization Strategies – ImplementationOptimization ResultsActivity 3.01: Optimizing a Deep Learning ModelSummary
IntroductionHandling New DataSeparating Data and ModelThe Data ComponentThe Model ComponentDealing with New DataRetraining an Old ModelTraining a New ModelExercise 4.01: Retraining a Model DynamicallyDeploying a Model as a Web ApplicationApplication Architecture and TechnologiesExercise 4.02: Deploying and Using CryptonicActivity 4.01: Deploying a Deep Learning ApplicationSummary
1. Introduction to Neural Networks and Deep LearningActivity 1.01: Training a Neural Network with Different Hyperparameters2. Real-World Deep Learning: Predicting the Price of BitcoinActivity 2.01: Assembling a Deep Learning System3. Real-World Deep Learning: Evaluating the Bitcoin ModelActivity 3.01: Optimizing a Deep Learning Model4. ProductizationActivity 4.01: Deploying a Deep Learning Application

Content preview from The Applied TensorFlow and Keras Workshop

The Applied TensorFlow and Keras Workshop

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 ...