TensorFlow 2.x in the Colaboratory Cloud : An Introduction to Deep Learning on Google’s Cloud Service

Intermediate-Advanced user levelUse TensorFlow 2.x with Google's Colaboratory (Colab) product that offers a free cloud service for Python programmers. Colab is especially well suited as a platform for TensorFlow 2.x deep learning applications. You will learn Colab's default install of the most current TensorFlow 2.x along with Colab's easy access to on-demand GPU hardware acceleration in the cloud for fast execution of deep learning models. This book offers you the opportunity to grasp deep learning in an applied manner with the only requirement being an Internet connection. Everything else--Python, TensorFlow 2.x, GPU support, and Jupyter Notebooks--is provided and ready to go from Colab. The book begins with an introduction to TensorFlow 2.x and the Google Colab cloud service. You will learn how to provision a workspace on Google Colab and build a simple neural network application. From there you will progress into TensorFlow datasets and building input pipelines in support of modeling and testing. You will find coverage of deep learning classification and regression, with clear code examples showing how to perform each of those functions. Advanced topics covered in the book include convolutional neural networks and recurrent neural networks.This book contains all the applied math and programming you need to master the content. Examples range from simple to relatively complex when necessary to ensure acquisition of appropriate deep learning concepts and constructs. Examples are carefully explained, concise, accurate, and complete to perfectly complement deep learning skill development. Care is taken to walk you through the foundational principles of deep learning through clear examples written in Python that you can try out and experiment with using Google Colab from the comfort of your own home or office. Table of Contents About the Author About the Technical Reviewer Introduction Chapter 1: Introduction to Deep Learning Neural Networks Learning Representation from Data TensorFlow 2.x Google Colab Google Drive Create a New Notebook GPU Hardware Accelerator Download a File from a URL Prepare the Dataset Colab Abends Colab Strange Results Tensors Scalars (0D Tensors) Vectors (1D Tensors) Matrices (2D Tensors) 3D Matrices (3D Tensors) Key Attributes of Tensors Input Pipelines The tf.data API Function from_tensor_slices Iterate a tf.data.Dataset Tensors and numpy Chaining Transformations Mapping Tensors Filter a tf.data.Dataset Shuffling a Dataset TensorFlow Math Vector Tensors Matrix Tensors tf.data.Dataset Tensors Save a Notebook Download a Notebook to a Local Drive Load a Notebook from a Local Drive Chapter 2: Build Your First Neural Network with Google Colab GPU Hardware Accelerator The load_digits Dataset Explore the Dataset Image Matrix Split Data into Train-Test Sets Build the Input Pipeline Explore TensorFlow Data Shuffle Data Continue Pipeline Construction Feedforward Neural Networks Number of Layers Our Model Model Summary Compile the Model Train the Model Model History Predictions Get an Image Mount Google Drive to Display an Image Chapter 3: Working with TensorFlow Data TensorFlow Datasets (TFDS) Colab Abends Available TFDS Load a TFDS Extract Useful Information Inspect the TFDS Feature Dictionaries Build the Input Pipeline Build the Model Model Summary Compile and Train the Model Generalize on Test Data Visualize Performance DatasetBuilder (tfds.builder) MNIST Metadata Show Examples Prepare DatasetBuilder Data Build the Model Compile the Model Train the Model Generalize on Test Data Load CIFAR-10 Inspect the Dataset Prepare the Input Pipeline Model the Data Chapter 4: Working with Other Data Basic Mechanics TensorFlow Dataset Structure Reading Input Data Colab Abends Batch Size Keras Data Build the Input Pipeline Create the Model Model Summary Compile the Model Train the Model Scikit-Learn Data Explore the Data Build the Input Pipeline Build the Model Model Summary Compile the Model Train the Model Numpy Data Load Numpy Arrays with tf.data.Dataset Prepare Data for Training Create the Model Model Summary Compile the Model Train the Model Get Wine Data from GitHub CSV Data Dataset Characteristics Get Data Split Data into Train and Test Sets Prepare Data for TensorFlow Consumption Build the Model Model Summary Compile the Model Train the Model Get Abalone Data from GitHub Data Datasets Abalone Dataset Dataset Characteristics Mount Google Drive to Colab Read Data Explore Data Create Train and Test Sets Create Feature and Target Sets Scale Features Create Train and Test Sets with Sex and Scaled Values Convert Numpy Feature Sets into Pandas DataFrames Build the Input Pipeline Explore a Batch Categorical Columns Build the Model Compile the Model Train the Model Imbalanced and Irregular Data Dealing with Imbalanced Data Chapter 5: Classification Fashion-MNIST Dataset Load Fashion-MNIST as a TFDS Explore the Dataset Build the Input Pipeline Build the Model Model Summary Compile the Model Train the Model Generalize on Test Data Visualize Performance Predict Labels for Test Images Build a Prediction Plot Load Fashion-MNIST as a Keras Dataset Explore the Data Visualize the First Image Visualize Sample Images Prepare Data for Training Build the Model Model Summary Compile the Model Train the Model Generalize on Test Data Visualize Training Predict Labels for Test Images Predict the First Image Predict Four Images Explore Misclassifications Visualize Misclassifications Predict from a Single Image Visualize Single Image Prediction Confusion Matrix Number of Hidden Layers Number of Neurons in Hidden Layers Chapter 6: Regression Boston Housing Dataset Boston Data Explore the Dataset Create Feature and Target Sets Get Feature Names from the Features DataFrame Convert Features and Labels Split Dataset into Train and Test Sets Scale Data and Create TensorFlow Tensors Prepare Tensors for Training Create a Model Model Summary Compile the Model Train the Model Visualize Training Early Stopping Remove Bad Data Get Data Remove Noise Create Feature and Target Data Build the Input Pipeline Compile and Train Visualize Generalize on Test Data Make Predictions Visualize Predictions Load Boston Data from Scikit-Learn Remove Noise Build the Input Pipeline Model Data Model Cars Data Get Cars Data from GitHub Convert Categorical Column to Numeric Slice Extraneous Data Create Features and Labels Build the Input Pipeline Model Data Inspect the Model Visualize Training Generalize on Test Data Make Predictions Visualize Predictions Chapter 7: Convolutional Neural Networks CNN Architecture Load Sample Images Display Images Scale Images Display Scaled Images Get More Images Mount Google Drive Copy Images to Google Drive Check Image Shapes Resize Images Create a Batch of Images Create Filters Plot Convolutional Kernels Apply a 2D Convolutional Layer Visualize Feature Maps CNN with Trainable Filters Building a CNN Load Data Display Information About the Dataset Extract Class Labels Display Samples Build a Custom Function to Display Samples Build a Custom Function to Display a Grid of Examples Pinpoint Metadata Build the Input Pipeline Create the Model Model Summary Model Layers Compile the Model Train the Model Generalize on Test Data Visualize Training Performance Predict Labels for Test Images Build a Prediction Plot Build a Custom Function Build a CNN with Keras Data Create Variables to Hold Train and Test Samples Display Sample Images Create the Input Pipeline Create the Model Compile and Train Predict Visualize Results Epilogue Chapter 8: Automated Text Generation Natural Language Processing Generating Text with a RNN The Text File Mount Google Drive Read the Corpus into Memory Verify Corpus Create Vocabulary Vectorize the Text Create Integer Mappings Create Character Mappings Map a Sequence Vectorize the Corpus Predict the Next Character Create Training Input Sequences Display Samples Batch Sequences Create Samples and Targets Time Step Prediction Create Training Batches Build the Model Display Model Summary Check Output Shape Calculate Loss Compile the Model Configure Checkpoints Train the Model Rebuild the Model for Text Creation Restore Weights from Checkpoints Rebuild with Batch Size of 1 Load Weights and Reshape Model Summary Create Components to Generate Text Create the Function Initialize Variables Vectorize and Reshape the Starting String Create New Text Chapter 9: Sentiment Analysis IMDb Dataset Load IMDb as a TFDS Display the Keys Split into Train and Test Sets Display the First Sample Display Information About the TFDS Peruse Metadata Create the Encoder Use the Encoder on Samples Finish the Input Pipeline Create the Model Model Summary Compile the Model Train the Model Generalize on Test Data Visualize Training Performance Make Predictions from Fabricated Reviews Make Predictions on a Test Data Batch Prediction Accuracy for the First Batch Leverage Pretrained Embeddings Load the IMDb Dataset Build the Input Pipeline Create the Pretrained Model Compile the Model Train the Model Make Predictions Calculate Prediction Accuracy for the First Batch Explore IMDb with Keras Explore the Train Sample Create a Decoding Function Invoke the Decoding Function Continue Exploring the Training Sample Train Keras IMDb Data Load Data Pad Samples Build the Input Pipeline Build the Model Compile the Model Train the Model Predict Chapter 10: Time Series Forecasting with RNNs Weather Forecasting The Weather Dataset Explore the Data Plot Relative Humidity Over Time Forecast a Univariate Time Series Scale Data Establish Training Split Create Features and Labels Create Train and Test Sets View Windows of Past History Plot a Single Example Create a Visual Performance Baseline Create a Baseline Metric Finish the Input Pipeline Explore a Data Window Create the Model Model Summary Verify Model Output Compile the Model Train the Model Generalize on Test Data Make Predictions Plot Model Performance Forecast a Multivariate Time Series Scale Data Multistep Model Generators Advantages of Using a Generator Generator Caveats Create a Generator Function Generate Train and Test Data Reconstitute Generated Tensors Finish the Input Pipeline Create the Model Model Summary Compile the Model Train the Model Generalize on Test Data Plot Performance Plot a Data Window Make a Prediction Index Use TensorFlow 2.x with Google's Colaboratory (Colab) product that offers a free cloud service for Python programmers. Colab is especially well suited as a platform for TensorFlow 2.x deep learning applications. You will learn Colab's default install of the most current TensorFlow 2.x along with Colab's easy access to on-demand GPU hardware acceleration in the cloud for fast execution of deep learning models. This book offers you the opportunity to grasp deep learning in an applied manner with the only requirement being an Internet connection. Everything else—Python, TensorFlow 2.x, GPU support, and Jupyter Notebooks—is provided and ready to go from Colab. The book begins with an introduction to TensorFlow 2.x and the Google Colab cloud service. You will learn how to provision a workspace on Google Colab and build a simple neural network application. From there you will progress into TensorFlow datasets and building input pipelines in support of modeling and testing. You will find coverage of deep learning classification and regression, with clear code examples showing how to perform each of those functions. Advanced topics covered in the book include convolutional neural networks and recurrent neural networks. This book contains all the applied math and programming you need to master the content. Examples range from simple to relatively complex when necessary to ensure acquisition of appropriate deep learning concepts and constructs. Examples are carefully explained, concise, accurate, and complete to perfectly complement deep learning skill development. Care is taken to walk you through the foundational principles of deep learning through clear examples written in Python that you can try out and experiment with using Google Colab from the comfort of your own home or office. What You Will Learn Be familiar with the basic concepts and constructs of applied deep learning Create machine learning models with clean and reliable Python code Work with datasets common to deep learning applications Prepare data for TensorFlow consumption Take advantage of Google Colab's built-in support for deep learning Execute deep learning experiments using a variety of neural network models Be able to mount Google Colab directly to your Google Drive account Visualize training versus test performance to see model fit Who This Book Is For Readers who want to learn the highly popular TensorFlow 2.x deep learning platform, those who wish to master deep learning fundamentals that are sometimes skipped over in the rush to be productive, and those looking to build competency with a modern cloud service tool such as Google Colab Use TensorFlow 2.x with Google's Colaboratory (Colab) product that offers a free cloud service for Python programmers. Colab is especially well suited as a platform for TensorFlow 2.x deep learning applications. You will learn Colabђ́ةs default install of the most current TensorFlow 2.x along with Colabђ́ةs easy access to on-demand GPU hardware acceleration in the cloud for fast execution of deep learning models. This book offers you the opportunity to grasp deep learning in an applied manner with the only requirement being an Internet connection. Everything elseђ́ؤPython, TensorFlow 2.x, GPU support, and Jupyter Notebooksђ́ؤis provided and ready to go from Colab. The book begins with an introduction to TensorFlow 2.x and the Google Colab cloud service. You will learn how to provision a workspace on Google Colab and build a simple neural network application. From there you will progress into TensorFlow datasets and building input pipelines in support of modeling and testing. You will find coverage of deep learning classification and regression, with clear code examples showing how to perform each of those functions. Advanced topics covered in the book include convolutional neural networks and recurrent neural networks. This book contains all the applied math and programming you need to master the content. Examples range from simple to relatively complex when necessary to ensure acquisition of appropriate deep learning concepts and constructs. Examples are carefully explained, concise, accurate, and complete to perfectly complement deep learning skill development. Care is taken to walk you through the foundational principles of deep learning through clear examples written in Python that you can try out and experiment with using Google Colab from the comfort of your own home or office. You will: Be familiar with the basic concepts and constructs of applied deep learning Create machine learning models with clean and reliable Python code Work with datasets common to deep learning applications Prepare data for TensorFlow consumption Take advantage of Google Colabђ́ةs built-in support for deep learning Execute deep learning experiments using a variety of neural network models Be able to mount Google Colab directly to your Google Drive account Visualize training versus test performance to see model fit