book

Generative AI with Python and TensorFlow 2

Name: Generative AI with Python and TensorFlow 2
ISBN: 9781800200883

by Joseph Babcock, Raghav Bali

April 2021

Intermediate to advanced

488 pages

12h 1m

English

Packt Publishing

Read now

Unlock full access

Preface
Who this book is forWhat this book coversTo get the most out of this bookDownload the example code filesDownload the color imagesConventions usedGet in touchReviews
An Introduction to Generative AI: "Drawing" Data from Models
Applications of AIDiscriminative and generative modelsImplementing generative modelsThe rules of probabilityDiscriminative and generative modeling and Bayes' theoremWhy use generative models?The promise of deep learningBuilding a better digit classifierGenerating imagesStyle transfer and image transformationFake news and chatbotsSound compositionThe rules of the gameUnique challenges of generative modelsSummaryReferences
Setting Up a TensorFlow Lab
Deep neural network development and TensorFlowTensorFlow 2.0VSCodeDocker: A lightweight virtualization solutionImportant Docker commands and syntaxConnecting Docker containers with docker-composeKubernetes: Robust management of multi-container applicationsImportant Kubernetes commandsKustomize for configuration managementKubeflow: an end-to-end machine learning labRunning Kubeflow locally with MiniKFInstalling Kubeflow in AWSInstalling Kubeflow in GCPInstalling Kubeflow on AzureInstalling Kubeflow using TerraformA brief tour of Kubeflow's componentsKubeflow notebook serversKubeflow pipelinesUsing Kubeflow Katib to optimize model hyperparametersSummaryReferences
Building Blocks of Deep Neural Networks
Perceptrons – a brain in a functionFrom tissues to TLUsFrom TLUs to tuning perceptronsMulti-layer perceptrons and backpropagationBackpropagation in practiceThe shortfalls of backpropagationVarieties of networks: Convolution and recursiveNetworks for seeing: Convolutional architecturesEarly CNNsAlexNet and other CNN innovationsAlexNet architectureNetworks for sequence dataRNNs and LSTMsBuilding a better optimizerGradient descent to ADAMXavier initializationSummaryReferences
Teaching Networks to Generate Digits
The MNIST databaseRetrieving and loading the MNIST dataset in TensorFlowRestricted Boltzmann Machines: generating pixels with statistical mechanicsHopfield networks and energy equations for neural networksModeling data with uncertainty with Restricted Boltzmann MachinesContrastive divergence: Approximating a gradientStacking Restricted Boltzmann Machines to generate images: the Deep Belief NetworkCreating an RBM using the TensorFlow Keras layers APICreating a DBN with the Keras Model APISummaryReferences
Painting Pictures with Neural Networks Using VAEs
Creating separable encodings of imagesThe variational objectiveThe reparameterization trickInverse Autoregressive FlowImporting CIFARCreating the network from TensorFlow 2 SummaryReferences
Image Generation with GANs
The taxonomy of generative modelsGenerative adversarial networksThe discriminator modelThe generator modelTraining GANsNon-saturating generator costMaximum likelihood gameVanilla GANImproved GANsDeep Convolutional GANVector arithmeticConditional GANWasserstein GANProgressive GANThe overall methodProgressive growth-smooth fade-inMinibatch standard deviationEqualized learning ratePixelwise normalizationTensorFlow Hub implementationChallengesTraining instabilityMode collapseUninformative loss and evaluation metricsSummaryReferences
Style Transfer with GANs
Paired style transfer using pix2pix GANThe U-Net generatorThe Patch-GAN discriminatorLossTraining pix2pixUse casesUnpaired style transfer using CycleGANOverall setup for CycleGANAdversarial lossCycle lossIdentity lossOverall lossHands-on: Unpaired style transfer with CycleGANGenerator setupDiscriminator setupGAN setupThe training loopRelated worksDiscoGANDualGANSummaryReferences
Deepfakes with GANs
Deepfakes overviewModes of operationReplacementRe-enactmentEditingKey feature setFacial Action Coding System (FACS)3D Morphable ModelFacial landmarksFacial landmark detection using OpenCVFacial landmark detection using dlibFacial landmark detection using MTCNNHigh-level workflowCommon architecturesEncoder-Decoder (ED)Generative Adversarial Networks (GANs)Replacement using autoencodersTask definitionDataset preparationAutoencoder architectureTraining our own face swapperResults and limitationsRe-enactment using pix2pixDataset preparationPix2pix GAN setup and trainingResults and limitationsChallengesEthical issuesTechnical challengesGeneralizationOcclusionsTemporal issuesOff-the-shelf implementationsSummaryReferences
The Rise of Methods for Text Generation
Representing textBag of WordsDistributed representationWord2vecGloVeFastTextText generation and the magic of LSTMsLanguage modelingHands-on: Character-level language modelDecoding strategiesGreedy decodingBeam searchSampling Hands-on: Decoding strategiesLSTM variants and convolutions for textStacked LSTMs Bidirectional LSTMsConvolutions and textSummaryReferences

NLP 2.0: Using Transformers to Generate Text
AttentionContextual embeddingsSelf-attentionTransformersOverall architectureMulti-head self-attentionPositional encodingsBERT-ologyGPT 1, 2, 3…Generative pre-training: GPTGPT-2Hands-on with GPT-2Mammoth GPT-3SummaryReferences
Composing Music with Generative Models
Getting started with music generationRepresenting musicMusic generation using LSTMsDataset preparationLSTM model for music generationMusic generation using GANsGenerator networkDiscriminator networkTraining and resultsMuseGAN – polyphonic music generationJamming modelComposer modelHybrid modelTemporal modelMuseGANGeneratorsCriticTraining and resultsSummaryReferences
Play Video Games with Generative AI: GAIL
Reinforcement learning: Actions, agents, spaces, policies, and rewardsDeep Q-learningInverse reinforcement learning: Learning from expertsAdversarial learning and imitationRunning GAIL on PyBullet GymThe agent: Actor-Critic networkThe discriminatorTraining and resultsSummaryReferences
Emerging Applications in Generative AI
Finding new drugs with generative modelsSearching chemical space with generative molecular graph networksFolding proteins with generative modelsSolving partial differential equations with generative modelingFew shot learning for creating videos from imagesGenerating recipes with deep learningSummaryReferences
Why subscribe?
Other Books You May Enjoy
Index

Overview

Dive into the exciting world of generative AI with 'Generative AI with Python and TensorFlow 2.' This book explores powerful AI techniques and gives you the tools to create amazing outputs like images, text, and music. With hands-on coding examples and real-world use cases, it's perfect for those ready to blend creativity with technical expertise.

What this Book will help me do

Understand and implement state-of-the-art generative models like GANs and VAEs
Develop practical skills in using TensorFlow 2 to build AI models
Create fascinating applications such as deepfakes, text generation, and music composition
Analyze cutting-edge research in generative AI to expand your knowledge
Apply transformative technologies like transformers in NLP projects

Author(s)

Joseph Babcock and Raghav Bali are expert practitioners in machine learning and AI with extensive experience in applying these technologies to solve real-world problems. Their clear and practical approach makes complex topics accessible to learners.

Who is it for?

This book is suited for Python programmers eager to explore generative AI techniques. If you have a basic grasp of machine learning concepts and are curious to learn about advanced AI applications and implementations, this book 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: 9781800200883

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills