40,81 €
Mehr erfahren.
- Herausgeber: Packt Publishing
- Kategorie: Wissenschaft und neue Technologien
- Sprache: Englisch
Tackle the complex challenges faced while building end-to-end deep learning models using modern R libraries
Key Features
- Understand the intricacies of R deep learning packages to perform a range of deep learning tasks
- Implement deep learning techniques and algorithms for real-world use cases
- Explore various state-of-the-art techniques for fine-tuning neural network models
Book Description
Deep learning (DL) has evolved in recent years with developments such as generative adversarial networks (GANs), variational autoencoders (VAEs), and deep reinforcement learning. This book will get you up and running with R 3.5.x to help you implement DL techniques.
The book starts with the various DL techniques that you can implement in your apps. A unique set of recipes will help you solve binomial and multinomial classification problems, and perform regression and hyperparameter optimization. To help you gain hands-on experience of concepts, the book features recipes for implementing convolutional neural networks (CNNs), recurrent neural networks (RNNs), and Long short-term memory (LSTMs) networks, as well as sequence-to-sequence models and reinforcement learning. You'll then learn about high-performance computation using GPUs, along with learning about parallel computation capabilities in R. Later, you'll explore libraries, such as MXNet, that are designed for GPU computing and state-of-the-art DL. Finally, you'll discover how to solve different problems in NLP, object detection, and action identification, before understanding how to use pre-trained models in DL apps.
By the end of this book, you'll have comprehensive knowledge of DL and DL packages, and be able to develop effective solutions for different DL problems.
What you will learn
- Work with different datasets for image classification using CNNs
- Apply transfer learning to solve complex computer vision problems
- Use RNNs and their variants such as LSTMs and Gated Recurrent Units (GRUs) for sequence data generation and classification
- Implement autoencoders for DL tasks such as dimensionality reduction, denoising, and image colorization
- Build deep generative models to create photorealistic images using GANs and VAEs
- Use MXNet to accelerate the training of DL models through distributed computing
Who this book is for
This deep learning book is for data scientists, machine learning practitioners, deep learning researchers and AI enthusiasts who want to learn key tasks in deep learning domains using a recipe-based approach. A strong understanding of machine learning and working knowledge of the R programming language is mandatory.
Das E-Book können Sie in Legimi-Apps oder einer beliebigen App lesen, die das folgende Format unterstützen:
Seitenzahl: 295
Veröffentlichungsjahr: 2020
Ähnliche
Deep Learning with R Cookbook
Copyright © 2020 Packt Publishing
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 express or implied. Neither the authors, nor Packt Publishing or its dealers and distributors, will be held liable for any damages caused or alleged to have been caused directly or indirectly by this book.
Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.
Commissioning Editor: Sunith ShettyAcquisition Editor: Yogesh DeokarContent Development Editor: Nathanya DiasSenior Editor: Ayaan HodaTechnical Editor: Joseph SunilCopy Editor: Safis EditingProject Coordinator: Aishwarya MohanProofreader: Safis EditingIndexer: Priyanka DhadkeProduction Designer: Jyoti Chauhan
First published: February 2020
Production reference: 1210220
Published by Packt Publishing Ltd. Livery Place 35 Livery Street Birmingham B3 2PB, UK.
ISBN 978-1-78980-567-3
www.packt.com
Packt.com
Subscribe to our online digital library for full access to over 7,000 books and videos, as well as industry-leading tools to help you plan your personal development and advance your career. For more information, please visit our website.
Why subscribe?
Spend less time learning and more time coding with practical eBooks and Videos from over 4,000 industry professionals
Improve your learning with Skill Plans built especially for you
Get a free eBook or video every month
Fully searchable for easy access to vital information
Copy and paste, print, and bookmark content
Did you know that Packt offers eBook versions of every book published, with PDF and ePub files available? You can upgrade to the eBook version at www.packt.com and as a print book customer, you are entitled to a discount on the eBook copy. Get in touch with us at [email protected] for more details.
At www.packt.com, you can also read a collection of free technical articles, sign up for a range of free newsletters, and receive exclusive discounts and offers on Packt books and eBooks.
Foreword
Data and AI give the best hope to the toughest problems that the world faces today. Am I making a sweeping statement? Not really—it's a modest statement of fact. From robotics to self-driving cars, farming that alleviates world hunger, to finding a solution to early diagnostics to critical illness—deep learning is one of the most enthralling areas of discovery and disruption. It has also fuelled the transformation of numerous businesses such as media and entertainment, insurance, healthcare, retail, education, and information technology.
This book is the perfect material for every data science enthusiast who wants to understand the concepts of deep learning: with R codes explained comprehensibly, it is the best place to start. The authors have maintained a perfect balance between theoretical and practical aspects of deep learning algorithms and applications. It turned out to be a great read—thanks to the easy flow of various sections such as Getting Ready, How to Do it, and How it Works. After starting with some good insights on how to set up a deep learning environment in a local system, the authors address how the reader can leverage various cloud platforms such as AWS, Microsoft Azure, and Google Cloud to scale deep learning applications. If you are looking for some quick thoughts on any topic, you can read any chapter individually without getting bogged about the sequence.
An interesting fact about this book is that it not only covers the generic topics of deep learning such as CNN, RNN, GAN, Autoencoders but also throws light on specific state-of-the-art techniques such as transfer learning and reinforcement learning. I like the practical examples in the chapters: Working with Convolutional Networks, Deep Generative models, Working with Text and Audio and NLP. They are bound to kindle some thought-starters on what can be done using image and text data. The data sets are very aptly chosen for the examples provided.
Overall, this book is an engaging and inspiring read. I congratulate the writers of the book- Swarna, Rehan, and Dipayan for their contribution to this field of study and I look forward to more such works from them.
Pradeep Jayaraman
Head of Analytics, Adani Ports & SEZ
Contributors
About the authors
Swarna Gupta holds a BE in computer science and has 6 years' experience in data science. She is currently working with Rolls Royce as a data scientist. Her work revolves around leveraging deep learning and machine learning to create value for the business. She has extensively worked on IoT-based projects in the vehicle telematics and solar manufacturing industries. During her current association with Rolls Royce, she implemented various deep learning techniques to build advanced analytics capabilities in the aerospace domain. Swarna also manages to find the time in her busy schedule to be a regular pro-bono contributor to social organizations, helping them to solve specific business problems with the help of data science and machine learning.
Rehan Ali Ansari has a BE in electrical and electronics engineering with 5 years' experience in data science. He is currently associated with digital competency at AP Moller Maersk Group in the capacity of a data scientist. Rehan has a diverse background of working across multiple domains including fashion and retail, IoT, the renewable energy sector, trade finance, and supply chain management. Rehan is a firm believer in the agile method for developing AI products and solutions. He holds a strong insight into the latest technologies in the field of data science. Outside of his busy schedule, Rehan manages to explore new areas in the field of robotics and AI.
Dipayan Sarkar holds an Masters in economics and has over 17 years' experience. He has won international challenges in predictive modeling and takes a keen interest in the mathematics behind machine learning techniques. Before opting to become an independent consultant and mentor in the data science and machine learning space with various organizations and educational institutions, he served as a senior data scientist with Fortune 500 companies in the U.S. and Europe. He is currently associated with the Great Lakes Institute of Management as a visiting faculty (analytics), and BML Munjal University as an adjunct faculty (analytics and machine learning). He has co-authored a book Ensemble Machine Learning with Python, available from Packt Publishing.
About the reviewer
Sray Agarwal has been working as a data scientist for the last 12 years and has gained experience in various domains, including BFSI, e-commerce, retail, telecommunications, hospitality, travel, education, real estate, and entertainment, among many other sectors. He is currently working for Publicis Sapient as a data scientist and is based out of London. His expertise lies in predictive modeling, forecasting, and advanced machine learning. He possesses a deep understanding of algorithms and advanced statistics. He has a background in management and economics and has attained an MSc-equivalent qualification in data science and analytics. He is also a SAS-Certified Predictive Modeler. His current area of interest is fair and explainable machine learning.
Packt is searching for authors like you
If you're interested in becoming an author for Packt, please visit authors.packtpub.com and apply today. We have worked with thousands of developers and tech professionals, just like you, to help them share their insight with the global tech community. You can make a general application, apply for a specific hot topic that we are recruiting an author for, or submit your own idea.
Table of Contents
Title Page
Copyright and Credits
Deep Learning with R Cookbook
Dedication
About Packt
Why subscribe?
Foreword
Contributors
About the authors
About the reviewer
Packt is searching for authors like you
Preface
Who this book is for
What this book covers
To get the most out of this book
Download the example code files
Download the color images
Conventions used
Sections
Getting ready
How to do it…
How it works…
There's more…
See also
Get in touch
Reviews
Understanding Neural Networks and Deep Neural Networks
Setting up the environment
Getting ready 
How to do it...
How it works...
There's more...
See also
Implementing neural networks with Keras
Sequential API
Getting ready
How to do it...
How it works...
There's more...
See also
Functional API
How to do it...
How it works...
There's more...
TensorFlow Estimator API
Getting ready
How to do it...
How it works...
There's more...
See also
TensorFlow Core API
Getting ready
How to do it...
How it works...
Implementing a single-layer neural network
Getting ready
How to do it...
How it works...
There's more...
See also
Training your first deep neural network
Getting ready
How to do it...
How it works...
There's more...
See also
Working with Convolutional Neural Networks
Introduction to convolutional operations
Getting ready
How to do it...
How it works...
There's more...
See also
Understanding strides and padding
How to do it...
How it works...
Getting familiar with pooling layers
Getting ready
How to do it...
How it works...
There's more...
See also
Implementing transfer learning
Getting ready
How to do it...
How it works...
There's more...
See also
Recurrent Neural Networks in Action
Sentiment classification using RNNs
Getting ready
How to do it...
How it works...
There's more...
See also
Text generation using LSTMs
Getting ready
How to do it...
How it works...
There's more...
See also
Time series forecasting using GRUs
Getting ready
How to do it...
How it works...
There's more...
See also
Implementing bidirectional recurrent neural networks
How to do it...
How it works...
There's more...
Implementing Autoencoders with Keras
Implementing vanilla autoencoders
Getting ready
How to do it...
How it works...
There's more...
Dimensionality reduction using autoencoders
Getting ready
How to do it...
How it works...
There's more...
Denoising autoencoders
Getting ready
How to do it...
How it works...
There's more...
See also
Changing black and white into color
Getting ready
How to do it...
How it works...
See also
Deep Generative Models
Generating images with GANs
Getting ready
How to do it...
How it works...
There's more...
See also
Implementing DCGANs
Getting ready
How to do it...
How it works...
There's more...
See also
Implementing variational autoencoders
Getting ready
How to do it...
How it works...
See also
Handling Big Data Using Large-Scale Deep Learning
Deep learning on Amazon Web Services
Getting ready
How to do it...
How it works...
Deep learning on Microsoft Azure
Getting ready
How to do it...
How it works...
There's more...
See also
Deep learning on Google Cloud Platform
Getting ready
How to do it...
How it works...
There's more...
Accelerating with MXNet
Getting ready
How to do it...
How it works...
There's more...
Implementing a deep neural network using MXNet 
Getting ready
How to do it...
How it works...
Forecasting with MXNet
Getting ready
How to do it...
How it works...
Working with Text and Audio for NLP
Neural machine translation
Getting ready
How to do it...
How it works...
There's more...
See also
Summarizing text using deep learning
Getting ready
How to do it...
How it works...
There's more...
See also
Speech recognition
Getting ready
How to do it...
How it works...
There's more...
Deep Learning for Computer Vision
Object localization
Getting ready
How to do it...
How it works...
There's more...
See also
Face recognition
Getting ready
How to do it...
How it works...
There's more...
See also
Implementing Reinforcement Learning
Model-based RL using MDPtoolbox
Getting ready
How to do it...
How it works...
There's more...
Model-free RL
Getting ready
How to do it...
How it works...
See also
Cliff walking using RL
Getting ready
How to do it...
How it works...
There's more...
Other Books You May Enjoy
Leave a review - let other readers know what you think
Preface
Deep learning has taken a huge step in recent years with developments including generative adversarial networks (GANs), variational autoencoders, and deep reinforcement learning. This book serves as a reference guide in R 3.x that will help you implement deep learning techniques.
This book walks you through various deep learning techniques that you can implement in your applications using R 3.x. A unique set of recipes will help you solve regression, binomial classification, and multinomial classification problems, and explores hyper-parameter optimization in detail. You will also go through recipes that implement convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory (LSTM) networks, sequence-to-sequence models, GANs, and reinforcement learning. You will learn about high-performance computation involving large datasets that utilize GPUs, along with parallel computation capabilities in R, and you will also get familiar with libraries such as MXNet, which is designed for efficient GPU computing and state-of-the-art deep learning. You will also learn how to solve common and not-so-common problems in NLP, such as object detection and action identification, and you will leverage pre-trained models in deep learning applications.
By the end of the book, you will have a logical understanding of deep learning and different deep learning packages and will be able to build the most appropriate solutions to your problems.
Who this book is for
This book is for data scientists, machine learning practitioners, deep learning researchers, and AI enthusiasts who want to learn key tasks in the deep learning domain using a recipe-based approach. You will implement deep learning techniques and algorithms in common and not-so-common challenges faced in research work or projects. A strong understanding of machine learning and a working knowledge of R is mandatory.
What this book covers
Chapter 1, Understanding Neural Networks and Deep Neural Networks, will show us how to set up a deep learning environment to train models. The readers are then introduced to neural networks, starting from how neural networks work, what hidden layers are, what backpropagation is, and what activation functions are. This chapter uses the keras library to demonstrate the recipes.
Chapter 2, Working with Convolutional Neural Networks, will show us CNNs and will explain how they can be used to train models for image recognition and natural language processing based tasks. This chapter also covers various hyperparameters and optimizers used with CNNs.
Chapter 3, Recurrent Neural Networks in Action, will show us the fundamentals of RNNs with real-life implementation examples. We will also introduce LSTMs and gated recurrent units (GRUs), an extension of RNNs, and take a detailed walk-through of LSTM hyper-parameters. In addition to this, readers will learn how to build a bi-directional RNN model using Keras.
Chapter 4, Implementing Autoencoders with Keras, will introduce the implementation of various types of autoencoders using the keras library as the backend. Readers will also learn about various applications of autoencoders, such as dimensionality reduction and image coloring.
Chapter 5, Deep Generative Models, will show us the architecture of another method of deep neural networks,generative adversarial networks (GANs). We will demonstrate how to train a GAN model comprising of two pitting nets—a generator and a discriminator. This chapter also covers the practical implementation of variational autoencoders and compares them with GANs.
Chapter 6, Handling Big Data Using Large-Scale Deep Learning, contains case studies on high-performance computation involving large datasets utilizing GPUs. Readers will also be introduced to the parallel computation capabilities in R and libraries such as MXNet, which is designed for efficient GPU computing and state-of-the-art deep learning.
Chapter 7, Working with Text and Audio for NLP, contains case studies on various topics involving sequence data, including natural language processing (NLP) and speech recognition. The readers will implement end-to-end deep learning algorithms using various deep learning libraries.
Chapter 8, Deep Learning for Computer Vision, will provide end-to-end case studies on object detection and face identification.
Chapter 9, Implementing Reinforcement Learning, will walk us through the concepts of reinforcement learning step by step. Readers will learn about various methods, such as Markov Decision Processes, Q-Learning, and experience replay, and implement these methods in R using examples. Readers will also implement an end-to-end reinforcement learning example using R packages such as MDPtoolbox and Reinforcementlearning.
To get the most out of this book
A good understanding of machine learning and strong knowledge of R is necessary for this book.
Download the example code files
You can download the example code files for this book from your account at www.packt.com. If you purchased this book elsewhere, you can visit www.packtpub.com/support and register to have the files emailed directly to you.
You can download the code files by following these steps:
Log in or register at
www.packt.com
.
Select the
Support
tab.
Click on
Code Downloads
.
Enter the name of the book in the
Search
box and follow the onscreen instructions.
Once the file is downloaded, please make sure that you unzip or extract the folder using the latest version of:
WinRAR/7-Zip for Windows
Zipeg/iZip/UnRarX for Mac
7-Zip/PeaZip for Linux
The code bundle for the book is also hosted on GitHub at https://github.com/PacktPublishing/Deep-Learning-with-R-Cookbook. In case there's an update to the code, it will be updated on the existing GitHub repository.
We also have other code bundles from our rich catalog of books and videos available at https://github.com/PacktPublishing/. Check them out!
Download the color images
We also provide a PDF file that has color images of the screenshots/diagrams used in this book. You can download it here: http://www.packtpub.com/sites/default/files/downloads/9781789805673_ColorImages.pdf.
Conventions used
There are a number of text conventions used throughout this book.
CodeInText: Indicates code words in text, database table names, folder names, filenames, file extensions, pathnames, dummy URLs, user input, and Twitter handles. Here is an example: "In step 1, we imported the fashion MNIST data using the dataset_fashion_mnist() function and checked the dimensions of its training and testing partitions."
A block of code is set as follows:
fashion <- dataset_fashion_mnist()x_train <- fashion$train$xy_train <- fashion$train$yx_test <- fashion$test$xy_test <- fashion$test$y
Bold: Indicates a new term, an important word, or words that you see onscreen. For example, words in menus or dialog boxes appear in the text like this. Here is an example: "Go to Anaconda Navigator from the Start menu."
Sections
In this book, you will find several headings that appear frequently (Getting ready, How to do it..., How it works..., There's more..., and See also).
To give clear instructions on how to complete a recipe, use these sections as follows:
Getting ready
This section tells you what to expect in the recipe and describes how to set up any software or any preliminary settings required for the recipe.
How to do it…
This section contains the steps required to follow the recipe.
How it works…
This section usually consists of a detailed explanation of what happened in the previous section.
There's more…
This section consists of additional information about the recipe in order to make you more knowledgeable about the recipe.
See also
This section provides helpful links to other useful information for the recipe.
Get in touch
Feedback from our readers is always welcome.
General feedback: If you have questions about any aspect of this book, mention the book title in the subject of your message and email us at [email protected].
Errata: Although we have taken every care to ensure the accuracy of our content, mistakes do happen. If you have found a mistake in this book, we would be grateful if you would report this to us. Please visit www.packtpub.com/support/errata, selecting your book, clicking on the Errata Submission Form link, and entering the details.
Piracy: If you come across any illegal copies of our works in any form on the Internet, we would be grateful if you would provide us with the location address or website name. Please contact us at [email protected] with a link to the material.
If you are interested in becoming an author: If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, please visit authors.packtpub.com.
Reviews
Please leave a review. Once you have read and used this book, why not leave a review on the site that you purchased it from? Potential readers can then see and use your unbiased opinion to make purchase decisions, we at Packt can understand what you think about our products, and our authors can see your feedback on their book. Thank you!
For more information about Packt, please visit packt.com.
Understanding Neural Networks and Deep Neural Networks
Deep learning has transformed many traditional businesses, such as web search, advertising, and many more. A major challenge with the traditional machine learning approaches is that we need to spend a considerable amount of time choosing the most appropriate feature selection process before modeling. Besides this, these traditional techniques operate with some level of human intervention and guidance. However, with deep learning algorithms, we can get rid of the overhead of explicit feature selection since it is taken care of by the models themselves. These deep learning algorithms are capable of modeling complex and non-linear relationships within the data. In this book, we'll introduce you to how to set up a deep learning ecosystem in R. Deep neural networks use sophisticated mathematical modeling techniques to process data in complex ways. In this book, we'll showcase the use of various deep learning libraries, such as keras and MXNet, so that you can utilize their enriched set of functions and capabilities in order to build and execute deep learning models, although we'll primarily focus on working with the keras library. These libraries come with CPU and GPU support and are user-friendly so that you can prototype deep learning models quickly.
In this chapter, we will demonstrate how to set up a deep learning environment in R. You will also get familiar with various TensorFlow APIs and how to implement a neural network using them. You will also learn how to tune the various parameters of a neural network and also gain an understanding of various activation functions and their usage for different types of problem statements.
In this chapter, we will cover the following recipes:
Setting up the environment
Implementing neural networks with Keras
TensorFlow Estimator API
TensorFlow Core API
Implementing a single-layer neural network
Training your first deep neural network
Setting up the environment
Before implementing a deep neural network, we need to set up our system and configure it so that we can apply a variety of deep learning techniques. This recipe assumes that you have the Anaconda distribution installed on your system.
Getting ready
Let's configure our system for deep learning. It is recommended that you create a deep learning environment in Anaconda. If you have an older version of R in the conda environment, you need to update your R version to 3.5.x or above.
You also need to install the CUDA and cuDNN libraries for GPU support. You can read more about the prerequisites at https://tensorflow.rstudio.com/tools/local_gpu.html#prerequisties.
Please note that if your system does not have NVIDIA graphics support, then GPU processing cannot be done.
How it works...
Keras and TensorFlow programs can be executed on both CPUs and GPUs, though these programs usually run faster on GPUs. If your system does not support an NVIDIA GPU, you only need to install the CPU version. However, if your system has an NVIDIA GPU that meets all the prerequisites and you need to run performance-critical applications, you should install the GPU version. To run the GPU version of TensorFlow, we need an NVIDIA GPU, and then we need to install a variety of software components (CUDA Toolkit v9.0, NVIDIA drivers, and cuDNN v7.0) on the system.
In steps1 to 3, we created a new conda environment with both the R and Python kernels installed. In steps4 and 5, we installed the keras library in the environment we created.
There's more...
The only supported installation method on Windows is conda. Therefore, you should install Anaconda 3.x for Windows before installing keras. The keras package uses the TensorFlow backend by default. If you want to switch to Theano or CNTK, call the use_backend() function after loading the keras library.
For the Theano backend, use the following command:
library(keras)use_backend("theano")
For the CNTK backend, use the following command:
library(keras)use_backend("cntk")
Now, your system is ready to train deep learning models.
See also
You can find out more about the GPU version installation of keras and its prerequisites here: https://tensorflow.rstudio.com/tools/local_gpu.html.
Implementing neural networks with Keras
TensorFlow is an open source software library developed by Google for numerical computation using data flow graphs. The R interface for TensorFlow is developed by RStudio, which provides an interface for three TensorFlow APIs:
Keras
Estimator
Core
The keras, tfestimators, and tensorflow packages provide R interfaces to the aforementioned APIs, respectively. Keras and Estimator are high-level APIs, while Core is a low-level API that offers full access to the core of TensorFlow. In this recipe, we will demonstrate how we can build and train deep learning models using Keras.
Keras is a high-level neural network API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano. The R interface for Keras uses TensorFlow as its default backend engine. The keras package provides an R interface for the TensorFlow Keras API. It lets you build deep learning models in two ways, sequential and functional, both of which will be described in the following sections.
Sequential API
Keras's Sequential API is straightforward to understand and implement. It lets us create a neural network linearly; that is, we can build a neural network layer-by-layer where we initialize a sequential model and then stack a series of hidden and output layers on it.
How it works...
In step 1, we initialized a sequential model by calling the keras_model_sequential() function. In the next step, we stacked hidden and output layers by using a series of layer functions. The layer_dense() function adds a densely-connected layer to the defined model. The first layer of the sequential model needs to know what input shape it should expect, so we passed a value to the input_shape argument of the first layer. In our case, the input shape was equal to the number of features in the dataset. When we add layers to the keras sequential model, the model object is modified in-place, and we do not need to assign the updated object back to the original. The keras object's behavior is unlike most R objects (R objects are typically immutable). For our model, we used the relu activation function. The layer_activation() function creates an activation layer that takes input from the preceding hidden layer and applies activation to the output of our previous hidden layer. We can also use different functions, such as leaky ReLU, softmax, and more (activation functions will be discussed in Implementing a single-layer neural network recipe). In the output layer of our model, no activation was applied.
We can also implement various activation functions for each layer by passing a value to the activation argument in the layer_dense() function instead of adding an activation layer explicitly. It applies the following operation:
output=activation(dot(input, kernel)+bias)
Here, the activation argument refers to the element-wise activation function that's passed, while the kernel is a weights matrix that's created by the layer. The bias is a bias vector that's produced by the layer.
To train a model, we need to configure the learning process. We did this in step 3 using the compile() function. In our training process, we applied a stochastic gradient descent optimizer to find the weights and biases that minimize our objective loss function; that is, the mean squared error. The metrics argument calculates the metric to be evaluated by the model during training and testing.
In step 4, we looked at the summary of the model; it showed us information about each layer, such as the shape of the output of each layer and the parameters of each layer.
In the last step, we trained our model for a fixed number of iterations on the dataset. Here, the epochs argument defines the number of iterations. The validation_split argument can take float values between 0 and 1. It specifies a fraction of the training data to be used as validation data. Finally, batch_size defines the number of samples that propagate through the network.
There's more...
Training a deep learning model is a time-consuming task. If training stops unexpectedly, we can lose a lot of our work. The keras library in R provides us with the functionality to save a model's progress during and after training. A saved model contains the weight values, the model's configuration, and the optimizer's configuration. If the training process is interrupted somehow, we can pick up training from there.
The following code block shows how we can save the model after training:
# Save modelmodel_sequential %>% save_model_hdf5("my_model.h5")
If we want to save the model after each iteration while training, we need to create a checkpoint object. To perform this task, we use the callback_model_checkpoint() function. The value of the filepath argument defines the name of the model that we want to save at the end of each iteration. For example, if filepath is {epoch:02d}-{val_loss:.2f}.hdf5
