Zero to Deep Learning

with Tensorflow 2
Download the First Chapter

Stop wasting your time

wrestling with incomplete and confusing tutorials

  • Why is it hard to learn deep learning?

    There's just so much to know before I can even get started!

  • Neural networks, convolution, recurrent neural nets?!?

    What the? How do I even know where to start?

  • It's hard to find any quality or complete blogs on building deep learning for production

    The Internet is full of incomplete blogs and poorly written code snippets?

  • Seems like everyone is using it

    while I don't know where to start and I don't want to get left behind!

  • Googling only takes you so far...

    There is seemingly no coherent resource for gluing all the pieces together!

  • Time is money

    don't waste it sifting through blogs and unhelpful academic tutorials

  • What the heck is a _____?

    The vocabulary is foreign, what is a perceptron? A matrix transposition? a hidden layer??!

  • How about using it in our application?

    How do we integrate deep learning into our applications?

  • How does it all fit together?

    and what do I do with a deep learning model?

  • Not hitting deadlines?

    I still have a job to do and stopping to learn deep learning will waste a lot of time!

Detailed Examples

The deep learning libraries are extensive.

The vocabulary, the syntax, the algorithms, are all incredibly complex and wide-ranging. What is the right package to use, the right library to import?

      1
      2
      3
      4
      5
      6
      7
      8
      9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      30
      31
      32
      from tensorflow.keras.models import Sequential
      from tensorflow.keras.layers import Dense, Conv2D, MaxPooling2D
      from tensorflow.keras.layers import Flatten, Activation, BatchNormalization
      
      
      model = Sequential([
          Conv2D(32, (3, 3), input_shape=(32, 32, 3)),
          MaxPooling2D(pool_size=(2, 2)),
          Activation('relu'),
      
          Conv2D(32, (3, 3)),
          MaxPooling2D(pool_size=(2, 2)),
          Activation('relu'),
          BatchNormalization(),
      
          Flatten(),
      
          Dense(128, activation='relu'),
          Dense(10, activation='softmax')
      ])
      
      
      model.compile(loss='categorical_crossentropy',
                    optimizer='rmsprop',
                    metrics=['accuracy'])
      
      
      model.fit(X_train, y_train,
                batch_size=128,
                epochs=5,
                validation_split=0.3)
      
      
      

    Ready to add deep learning to your toolbelt?

    What if you knew exactly how all the pieces of a deep learning system work together and had a solid understanding of the mathematics where it runs - in less time, without banging your head against the wall. Imagine how quickly you could work if you knew the best practices and the right tools?

    Stop wasting your time searching and have everything you need to be productive in one, well-organized place, with complete examples to get your project running intelligently without needing to resort to endless hours of research.

    You will learn the right way to integrate deep learning with the latest tools with your current applications and even build some new ones along the way.

    Download the first chapter

    Interactive code

    Jupyter notebooks included

    Every chapter in the book comes with a complete interactive Jupyter notebook that uses the concepts in the chapter.

    Data walk-through

    Zero to Deep Learning starts out with a deep introduction into data manipulation. We aim to support both the season-professional and the complete beginner.

    We explore sample datasets statistics, work with pivot tables, and visualizing data patterns.

    Gentle introduction

    Zero to Deep Learning is specifically crafted to make deep learning accessible to web developers of all experience levels.

    Wide-ranging topics

    Zero to Deep Learning covers a wide-ranging number of topics, from image recognition through text detection. It covers production-level model serving to make it easy to apply deep learning to web applications today.

    Gradient Descent

    Zero to Deep Learning gently introduces deep learning topics with introductory topics, such as Gradient Descent before diving too far deeply into the deep-end.

    Convolutional Neural Networks

    A course on Deep Learning would be incomplete without a course on convolutional neural networks, the quitessential example of the power of deep learning.

    We work with powerful image recognition systems using convolutional neural networks, from the basics through the end-to-end system.

    Book Contents

    Zero to Deep Learning is carefully designed to teach you step-by-step how to build, train, evaluate, improve and deploy deep learning models. Each chapter covers a topic and we provide full code examples as executable Jupyter notebooks.

    • Getting Started1
    • Data Manipulation2
    • Machine Learning3
    • Deep Learning4
    • Deep Learning Internals5
    • Convolutional Neural Networks6
    • Time Series and Recurrent Neural Networks7
    • Natural Language Processing and Text Data8
    • Training with GPUs9
    • Performance Improvement10
    • Pre-trained Models for Images11
    • Pre-trained Models for Text12
    • Serving Deep Learning Models13
    • Conclusion and Next Steps14

    Get up and running quickly

    Within the first few minutes, we'll know enough deep learning to start seeing the benefits of using it in our applications.

    Interactive

    Every single chapter and line of code includes an interactive Jupyter notebook. You'll get access to a Jupyter notebook for all code samples.

    Best practices

    Learn the best practices, such as: handling overfitting, code organization, and how to serve our model to our apps. We'll walk through practical, common examples of how to implement complete applications powered by deep learning.

    Comprehensive topics

    You'll learn core deep learning concepts - from the multiperceptron through deep neural networks including convolutional and recurrent neural networks.

    App

    Learn by Example

    When you get Zero to Deep Learning, you're not buying just a book, but an interactive course with hundreds of code examples.

    Interactive code

    • Interactive Jupyter notebooks
    • Visually-driven code, explained step-by-step
    • Multiple exercises with every chapter

    750+ Pages

    A ton of content is included covering the very basics through the latest technical implementations of deep neural networks.

    Thousands of lines of code

    Using real-world examples, you'll have a TON of code to use to learn from.

    Interactive Jupyter notebooks

    The book includes runnable code examples for building all the code in the book.

    Too good to be true?

    Grab a sample chapter and check it out for yourself. Sign up for our mailing list and get the sample chapters for free! You'll only receive email about the book and updates. We never send spam, ever and it's easy to unsubscribe.

    Sample chapter image

    It can take up to an hour to deliver the sample chapter. If you don't receive the sample chapter within the hour, write us and we'll send them to you directly.

    Production-quality

    Built on Tensorflow 2 and Keras

    Learn practical Tensorflow applications with Keras – without getting lost in equations.

    With Zero to Deep Learning, you'll be armed with production-quality knowledge to take your new deep learning skills to professional products.

    Tensorflow 2

    An open-source software library for Machine Intelligence built and maintained by the brilliant engineers at Google.

    tf.keras

    Keras is a high-level neural networks API, written in Python and it was chosen as the default API for Tensorflow 2.0.

    Packages

    Grab a copy now

    Most Popular

    Book, source code, videos

    • Zero to Deep Learning Book
    • Completely DRM-free PDF, mobi, and epub formatted files
    • 14 Runnable, Interactive Jupyter Notebooks
    • Complete code for the examples in the book
    • Solutions to all the exercises
    • A screencast for advanced students on how to deploy models
    • -
    Get it now
    Team

    Team License

    • Zero to Deep Learning Book
    • Completely DRM-free PDF, mobi, and epub formatted files
    • 14 Runnable, Interactive Jupyter Notebooks
    • Complete code for the examples in the book
    • Solutions to all the exercises
    • Everything in 'Most Popular'
    • Team license for up to 10 team members
    • Immediate invoice billing service
    • Access to older versions of the book
    • Access to full Git repository of book and code
    • -
    Get it now

    The Team

    Meet the authors

    Francesco Mosconi

    Francesco Mosconi

    Francesco Mosconi is CEO and Chief Data Scientist at Catalit Data Science. He holds a joint Ph.D. in Physics and Biology. With Catalit, Francesco has been teaching immersive week-long bootcamps on Deep Learning and Data Science. Along the way, he has taught hundreds of developers, just like you, how to understand -- and how to use -- deep learning in their applications.

    Ari Lerner

    Ari Lerner

    Ari is a web development teacher and author of ng-book.

    Nate Murray

    Nate Murray

    Nate is a full-stack developer and writes code for everything from deep-learning image recognition to mobile games for cats. Nate formerly worked at IFTTT and his background is in data mining and scaling web services.

    FAQs

    Questions? We have answers!

    How long is the book?

    The final version has 14 chapters totaling 750+ pages, with runnable examples consisting of thousands of lines of code

    Do I have to know or be good at math?

    Nope! We don't assume that you're a math wizard. Instead, we take the approach of adding what you need to know when you need to know it. We also highlight the more mathematical sections that can be reviewed later.

    Are there free updates?

    Yes! Updates are free for 12-months following purchase. We've released over 50 updates to ng-book already

    What about Machine Learning?

    Chapter 3 contains a brief overview of Machine Learning, so you will get a chance to review its main concepts.

    Is this a physical or digital book?

    This is a completely DRM-free ebook formatted as a pdf/mobi/epub (and a zip with tons of example code)

    Is there a physical print version of the book?

    Soon! Q3-2019

    Does this cover Tensorflow 2.0?

    Yes!

    What if I don't like it?

    If you're unhappy with the book or content, just reach out to us and we'll give you a full refund. There's no risk.

    Our Promise to You

    We're committed to keeping Zero to Deep Learning as the best resource for learning and implementing deep learning into our applications. We personally respond to requests for content and we regularly release updates. We're independent authors and we survive by making the highest quality book on deep learning as possible.

    There's no risk: if you're not satisfied for any reason, send us an email and we'll give you a full refund.

    Contact Us

    If you have any concerns, feel free to email us