packtpublishing/mastering-keras

Mastering Keras, Published by Packt

Mastering Keras [Video]

This is the code repository for Mastering Keras [Video], published by Packt. It contains all the supporting project files necessary to work through the video course from start to finish.

About the Video Course

Successful data scientists need to be able to work with the most powerful tools to solve the most challenging problems. As deep learning becomes ever more entrenched as the gold-standard tool for a wide variety of advanced data analytics and Artificial Intelligence problems, it is essential for you as a data scientist or analyst to be comfortable wielding these powerful techniques on an ever-expanding array of problems.
TensorFlow (and its easy-to-learn deep learning wrapper Keras) have become game-changers in permitting simple implementations of the most complex of deep learning techniques.
In this course, we teach you to go beyond your working knowledge of Keras, begin to wield its full power, and unleash the amazing potential of advanced deep learning on your data science problems. You'll learn to design and train deep learning models for synthetic data generation, object detection, one-shot learning, and much more.
By the end of this course, you will be able to implement many advanced deep learning modelling algorithms and adapt them to your own purposes. Perhaps the next great breakthrough will come from you? Please note that familiarity with machine learning and deep learning approaches, together with practical experience with Keras and Python programming, are assumed for taking this course.

What You Will Learn

• Use the powerfully functional Keras API to design and implement advanced deep learning techniques
• Design and implement advanced Convolutional Neural Networks for powerful image classification
• Design and implement object detection networks to identify objects present in images and their location
• Work with deep generative neural networks for synthetic data generation and semi-supervised learning
• Develop a stable deep reinforcement-learning system and learn to make optimal decisions via feedback from their environment
• Implement deep one-shot learning systems that can classify new instances of a class after a single exposure to such an object

Instructions and Navigation

Assumed Knowledge

If you are a machine learning or deep learning professional (or enthusiast) and want to master advanced new techniques using Keras, then this course is for you!

Familiarity with machine learning approaches and practical experience with Keras are assumed. Fluency with Python programming is assumed.

Technical Requirements

Hardware Requirements

For successful completion of this course, we recommend a medium configuration computer. For the speedy execution of some code, it will be advantageous to have an NVIDIA GPU and be utilizing the GPU accelerated version of TensorFlow.

Software Requirements

The videos in this course include coding in Python Jupyter Notebook environment and make use of the Keras-Tensorflow toolchain. This ecosystem is managed using the Anaconda platform. We recommend that you use this platform, which is available at https://www.anaconda.com/:

If you wish to run Python and Jupyter Notebook without Anaconda, they can be installed individually:

1. Python - available at https://www.python.org/. You will also need to have the following libraries installed:
   • Keras version 2.24 or later
   • Tensorflow version 1.8 or later
   
   

If you want to use an NVIDIA GPU make sure you install the appropriate GPU versions of TensorFlow.

2. Jupyter Notebook - available at https://jupyter.org/
   
   

Related Products

• Mastering GitLab [Video]

• Mastering Palo Alto Networks [Video]

• Azure Cognitive Services for Developers [Video]

mastering-keras's Issues

Alice.txt is not found

Wrong label names on module 2.2

On module 2.2, history['acc'] and history['val_acc'] are called instead history['accuracy'] and history['val_accuracy']

def plot_training_history(history):
    plt.plot(history.history['acc'])
    plt.plot(history.history['val_acc'])

def plot_training_history(history):
    plt.plot(history.history['accuracy'])
    plt.plot(history.history['val_accuracy'])