book

The Deep Learning Workshop

Name: The Deep Learning Workshop
ISBN: 9781839219856

by Mirza Rahim Baig, Thomas Joseph, Nipun Sadvilkar, Mohan Kumar Silaparasetty, Anthony So, Akshay Chauhan, Nagendra Nagaraj, Robert Ridley

July 2020

Intermediate to advanced

474 pages

8h 51m

English

Packt Publishing

Read now

Unlock full access

The Deep Learning Workshop
Preface
About the BookAudienceAbout the ChaptersConventionsCode PresentationSetting up Your EnvironmentHardware RequirementsInstalling Anaconda on your systemLaunching Jupyter NotebookInstalling LibrariesInstalling TensorFlow 2.0Installing KerasAccessing the Code Files
1. Building Blocks of Deep Learning
IntroductionAI, Machine Learning, and Deep LearningMachine LearningDeep LearningUsing Deep Learning to Classify an ImagePre-Trained ModelsThe Google Text-to-Speech APIPrerequisite Packages for the DemoExercise 1.01: Image and Speech Recognition DemoDeep Learning ModelsThe Multi-Layer PerceptronConvolutional Neural NetworksRecurrent Neural NetworksGenerative Adversarial NetworksIntroduction to TensorFlowConstantsVariablesDefining Functions in TensorFlowExercise 1.02: Implementing a Mathematical EquationLinear Algebra with TensorFlowExercise 1.03: Matrix Multiplication Using TensorFlowThe reshape FunctionExercise 1.04: Reshaping Matrices Using the reshape() Function in TensorFlowThe argmax FunctionExercise 1.05: Implementing the argmax() FunctionOptimizersExercise 1.06: Using an Optimizer for a Simple Linear RegressionActivity 1.01: Solving a Quadratic Equation Using an OptimizerSummary
2. Neural Networks
IntroductionNeural Networks and the Structure of PerceptronsInput LayerWeightsBiasNet Input FunctionActivation Function (G)Perceptrons in TensorFlowExercise 2.01: Perceptron ImplementationTraining a PerceptronPerceptron Training Process in TensorFlowExercise 2.02: Perceptron as a Binary ClassifierMulticlass ClassifierThe Softmax Activation FunctionExercise 2.03: Multiclass Classification Using a PerceptronMNIST Case StudyExercise 2.04: Classifying Handwritten DigitsKeras as a High-Level APIExercise 2.05: Binary Classification Using KerasMultilayer Neural Network or Deep Neural NetworkReLU Activation FunctionExercise 2.06: Multilayer Binary ClassifierExercise 2.07: Deep Neural Network on MNIST Using KerasExploring the Optimizers and Hyperparameters of Neural NetworksGradient Descent OptimizersThe Vanishing Gradient ProblemHyperparameter TuningOverfitting and DropoutActivity 2.01: Build a Multilayer Neural Network to Classify Sonar SignalsSummary
3. Image Classification with Convolutional Neural Networks (CNNs)
IntroductionDigital ImagesImage ProcessingConvolution OperationsExercise 3.01: Implementing a Convolution OperationStridePaddingConvolutional Neural NetworksPooling LayersCNNs with TensorFlow and Keras Exercise 3.02: Recognizing Handwritten Digits (MNIST) with CNN Using KERAS Data GeneratorExercise 3.03: Classifying Cats versus Dogs with Data GeneratorsData AugmentationHorizontal FlippingVertical FlippingZoomingHorizontal ShiftingVertical ShiftingRotatingShearingExercise 3.04: Image Classification (CIFAR-10) with Data AugmentationActivity 3.01: Building a Multiclass Classifier Based on the Fashion MNIST DatasetSaving and Restoring ModelsSaving the Entire ModelSaving the Architecture OnlySaving the Weights OnlyTransfer LearningFine-TuningActivity 3.02: Fruit Classification with Transfer LearningSummary
4. Deep Learning for Text – Embeddings
IntroductionDeep Learning for Natural Language ProcessingGetting Started with Text Data HandlingText PreprocessingTokenizationNormalizing CaseRemoving PunctuationRemoving Stop WordsExercise 4.01: Tokenizing, Case Normalization, Punctuation, and Stop Word RemovalStemming and LemmatizationExercise 4.02: Stemming Our DataBeyond Stemming and LemmatizationDownloading Text Corpora Using NLTKActivity 4.01: Text Preprocessing of the 'Alice in Wonderland' TextText Representation ConsiderationsClassical Approaches to Text RepresentationOne-Hot EncodingExercise 4.03: Creating One-Hot Encoding for Our DataTerm FrequenciesThe TF-IDF MethodExercise 4.04: Document-Term Matrix with TF-IDFSummarizing the Classical ApproachesDistributed Representation for TextWord Embeddings and Word Vectorsword2vecTraining Our Own Word EmbeddingsSemantic Regularities in Word EmbeddingsExercise 4.05: Vectors for PhrasesEffect of Parameters – "size" of the VectorEffect of Parameters – "window size"Skip-gram versus CBOWEffect of Training DataExercise 4.06: Training Word Vectors on Different DatasetsUsing Pre-Trained Word VectorsBias in Embeddings – A Word of CautionOther Notable Approaches to Word EmbeddingsActivity 4.02: Text Representation for Alice in WonderlandSummary
5. Deep Learning for Sequences
IntroductionWorking with SequencesTime Series Data – Stock Price PredictionExercise 5.01: Visualizing Our Time-Series DataRecurrent Neural NetworksLoops – An Integral Part of RNNsExercise 5.02: Implementing the Forward Pass of a Simple RNN Using TensorFlowThe Flexibility and Versatility of RNNsPreparing the Data for Stock Price PredictionParameters in an RNNTraining RNNsExercise 5.03: Building Our First Plain RNN ModelModel Training and Performance Evaluation1D Convolutions for Sequence ProcessingExercise 5.04: Building a 1D Convolution-Based ModelPerformance of 1D ConvnetsUsing 1D Convnets with RNNsExercise 5.05: Building a Hybrid (1D Convolution and RNN) ModelActivity 5.01: Using a Plain RNN Model to Predict IBM Stock PricesSummary
6. LSTMs, GRUs, and Advanced RNNs
IntroductionLong-Range Dependence/InfluenceThe Vanishing Gradient ProblemSequence Models for Text ClassificationLoading DataStaging and Preprocessing Our DataThe Embedding LayerBuilding the Plain RNN ModelExercise 6.01: Building and Training an RNN Model for Sentiment ClassificationMaking Predictions on Unseen DataLSTMs, GRUs, and Other VariantsLSTMsParameters in an LSTMExercise 6.02: LSTM-Based Sentiment Classification ModelLSTM versus Plain RNNsGated Recurrence UnitsExercise 6.03: GRU-Based Sentiment Classification ModelLSTM versus GRUBidirectional RNNsExercise 6.04: Bidirectional LSTM-Based Sentiment Classification ModelStacked RNNsExercise 6.05: Stacked LSTM-Based Sentiment Classification ModelSummarizing All the ModelsAttention ModelsMore Variants of RNNsActivity 6.01: Sentiment Analysis of Amazon Product ReviewsSummary
7. Generative Adversarial Networks
IntroductionKey Components of Generative Adversarial NetworksProblem Statement – Generating a Distribution Similar to a Given Mathematical FunctionProcess 1 – Generating Real Data from the Known FunctionExercise 7.01: Generating a Data Distribution from a Known FunctionProcess 2 – Creating a Basic Generative NetworkBuilding the Generative NetworkSequential()Kernel InitializersDense LayersActivation FunctionsExercise 7.02: Building a Generative NetworkSetting the Stage for the Discriminator NetworkProcess 3 – Discriminator NetworkImplementing the Discriminator NetworkFunction to Generate Real SamplesFunctions to Generate Fake SamplesBuilding the Discriminator NetworkTraining the Discriminator NetworkExercise 7.03: Implementing the Discriminator NetworkProcess 4 – Implementing the GANIntegrating All the Building BlocksProcess for Building the GAN The Training ProcessExercise 7.04: Implementing the GANDeep Convolutional GANsBuilding Blocks of DCGANsGenerating Handwritten Images Using DCGANsThe Training ProcessExercise 7.05: Implementing the DCGANAnalysis of Sample PlotsCommon Problems with GANsMode CollapseConvergence FailureActivity 7.01: Implementing a DCGAN for the MNIST Fashion DatasetSummary
Appendix
1. Building Blocks of Deep LearningActivity 1.01: Solving a Quadratic Equation Using an OptimizerSolution2. Neural NetworksActivity 2.01: Build a Multilayer Neural Network to Classify Sonar SignalsSolution 3. Image Classification with Convolutional Neural Networks (CNNs)Activity 3.01: Building a Multiclass Classifier Based on the Fashion MNIST DatasetSolution Activity 3.02: Fruit Classification with Transfer LearningSolution 4. Deep Learning for Text – EmbeddingsActivity 4.01: Text Preprocessing of the 'Alice in Wonderland' TextSolutionActivity 4.02: Text Representation for Alice in WonderlandSolution5. Deep Learning for SequencesActivity 5.01: Using a Plain RNN Model to Predict IBM Stock PricesSolution 6. LSTMs, GRUs, and Advanced RNNsActivity 6.01: Sentiment Analysis of Amazon Product ReviewsSolution7. Generative Adversarial NetworksActivity 7.01: Implementing a DCGAN for the MNIST Fashion DatasetSolution

Overview

The Deep Learning Workshop is a hands-on guide designed to teach you the principles of deep learning using TensorFlow and Keras. Through practical exercises and real-world examples, you'll explore areas such as neural networks, CNNs, RNNs, and GANs, and learn how to implement these concepts to create powerful AI applications and models.

What this Book will help me do

Understand and distinguish between deep learning, machine learning, and AI.
Develop deep neural network architectures with TensorFlow.
Use Keras to train and enhance models for real-world applications.
Implement convolutional neural networks for image recognition tasks.
Learn to build and generate synthetic data using GANs.

Author(s)

The authors of The Deep Learning Workshop are seasoned deep learning professionals with vast teaching and practical experience. They are proficient in TensorFlow and Keras and aim to simplify complex topics through clear explanations and engaging exercises. Their collaborative approach ensures that concepts are conveyed effectively and inclusively.

Who is it for?

This book is ideal for tech enthusiasts looking to delve into deep learning. Whether you're a data scientist wanting to upskill in AI, a Python expert seeking to explore new ML frameworks, or someone with foundational ML knowledge aiming to learn TensorFlow and Keras, The Deep Learning Workshop is tailored for you.

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: 9781839219856

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills