A Practical Guide to Deep Learning with Keras

Name: A Practical Guide to Deep Learning with Keras
Rating: 4.5 (4 reviews)

Implement AI with Keras for building complex Deep Learning neural networks with fewer lines of coding in Python

Created byPackt Publishing

Last updated 7/2019

English

What you'll learn

Install and configure Keras. Study Deep Convolutional Neural Networks.
Develop a deep learning network from scratch with Keras using Python to solve a practical problem of classifying the traffic signs on the road.
Get introduced to Computer Vision & Deep Learning.
Setup and develop an environment with VM or Docker. Ipython and Jupyter notebook.
Discover activation functions, forward propagation, backward propagation.
Tensorboard and intuitions of filters and hyper-parameters.
Deploy and evaluate for other real-world applications. Future work and readings!
Learn Neural network style transfer - Image style translation and generation.
Develop Game AI - Running game agents using Deep Q network.

Course content

2 sections • 29 lectures • 3h 52m total length

The Course Overview3:09
This video gives an overview of the entire course.
Perceptron4:13
Perceptron is the name given to a model having one single linear layer, and as a consequence, if it has multiple layers, you would call it multilayer perceptron (MLP).
Understand what perceptron is
Look at the activation functions
Building a Network to Recognize Handwritten Numbers18:55
In this video we will be learning how to build neural network using Keras.
Defining a simple neural net in Keras
Run a simple Keras net and establishing a baseline
Improve the simple net in Keras with hidden layers
Improve the simple net in Keras with dropout
Playing Around with the Parameters to Improve Performance10:07
Here we look at many tunable parameters that can be tweaked to get good results from our network.
Test different optimizers in Keras
Control the optimizer learning rate
Learn about the Hyperparameters tuning
Installing and Configuring Keras9:58
In this video we will see how to install Keras on multiple platforms. Also, we will configure Keras, it has a very minimalist configuration file.
Install Keras
Configure Keras
Install Keras on Docker
Keras API5:49
Keras has a modular, minimalist, and easy extendable architecture. Here we will review the most important Keras components used for defining neural networks.
Look at the Keras architecture
See sequential composition
Understand functional composition
Callbacks for Customizing the Training Process3:22
The training process can be stopped when a metric has stopped improving by using an appropriate callback. We will also look at the Checkpointing.
Save lost history
Understand checkpointing
Use TesnsorBoard and Keras
Deep Convolutional Neural Network – DCNN11:37
A deep convolutional neural network (DCNN) consists of many neural network layers. Two different types of layers, convolutional and pooling, are typically alternated. The depth of each filter increases from left to right in the network. The last stage is typically made of one or more fully connected layers.
See the shared weights and bias
Understand pooling layers
Look at the LeNet code in Keras
Recognizing CIFAR-10 Images with Deep Learning13:59
The CIFAR-10 dataset contains 60,000 color images of 32 x 32 pixels in 3 channels divided into 10 classes. Each class contains 6,000 images. The training set contains 50,000 images, while the test sets provides 10,000 images.
Recognize previously unseen images and assign them to one of the 10 classes
Improve the CIFAR-10 performance with deeper a network
Improve the CIFAR-10 performance with data augmentation

The Course Overview2:57
This video will give you an overview about the course.
AI, Machine Learning, Deep Learning Overview3:55
In this video, you’ll be introduced to Deep Learning terminology and explain where Keras fits into the picture of Deep learning models.
Introduce AI, Machine learning, Deep learning terminologies
Understand Tensor in TensorFlow and its general naming implications
Understand where Keras fits in the picture
Keras Overview3:42
This video will give a brief introduction to Keras and explain why it is important.
Comparative analysis of different Deep Learning frameworks
Understand the different layers and core models in Keras briefly
Code simplicity analysis of Keras vs Tensorflow to understand the usability of Keras
Development Environment Setup4:39
In this video, we will set up the development environment for the course.
Use Anaconda on Ubuntu 14 or higher
Clone the Git repo and follow the environment setup commands
conda env create <> conda activate <> conda env list
Deep Learning(DL)8:46
This video will give you a brief understanding of Deep Learning. Develop an intuition of higher dimensional space modeling using neural networks
Intuition of Deep Learning and how it relates to modeling
Understand Machine learning ideas and concepts
Develop a good intuition of the field
Deep Neural Nets : Forward and Backward Propagation13:56
This video will introduce you to the deep neural network.
Understand Forward propagation
Understand Backward propagation
CNN Intuition12:00
This video introduces CNN and intuition of convolutions and Covnets.
Introduce convolutions, Convolution Maths and operations
Find out features from a dataset and strive towards minimizing error using pre-built Keras libraries
Explain where Keras fits in the picture
Autonomous Driving Simulator5:29
This video gives a brief introduction to autonomous driving simulators and figuring out how to drive autonomously. Which features to look at? Develop the model and compile it.
Download the simulator and drive autonomously
Understand the Keras code for model development
Build the model and visualize the loss function
Running in Autonomous Mode16:59
This video will give a brief on running the frozen model in autonomous mode and verify how it performs.
Run the frozen model using the command to help to drive autonomously
Retrain if necessary to go round the lap
Try other tracks
Reinforcement Learning and Q-Learning for Game AI7:16
This video will give you a brief understanding of Reinforcement Learning. Develop an intuition.
Gives an intuition of Reinforcement Learning and how it relates to modeling
Define Agent, Policy, Reward
Develop a good intuition of the field
Setting up Game Environment – OpenAI Gym with Keras6:55
This video will give you a brief introduction and intuition of OpenAI Gym.
Set up environment
Understand RL benchmarking
Learn RL training with Keras
Developing DQN-Based Dino Run Game9:20
In this video, you will get a brief introduction to a game environment and how to extract State, Action, and Reward out of it. Finally, train to run the agent in autonomous mode.
Download and setup everything
Understand the different parts of the code
Build the model and give intuition of changing the model to Keras
Developing Model for Dino Run Game to Run in AI Mode5:17
In this video, you will run the frozen model in autonomous mode and verify how it performs.
Run the frozen model using the command to help navigate autonomously
Retrain if necessary to score even higher
Try other tracks
Neural Network Style Transfer6:04
This video will give a brief understanding of the Style transfer algorithm using standard CNN.
Intuition of Style transfer
Understand the Loss function in 2 different styled images
Develop and understand the use of standard state-of-the-art DNN models
Neural Network Style Transfer – CNN Code Using Keras10:20
In this video, you will understand the code.
Understand TensorFlow eager execution
Display of results where the style transfer is happening
Understand the code execution in Jupyter notebook environment
Generative Adversarial Network9:17
In this video, we will look at different paradigm of NN – GAN.
Understand the loss function in GAN
Go from GAN to CGAN paradigm
Style transfer and generating different images using CGAN
Style Transfer – CGAN Code8:58
In this video, you will look at Style transfer and generating different images using CGAN
Example of a Style transfer code in CGAN
Develop the generator and discriminator
Derive and code the loss function
CNN Problem and Trends6:24
In this video, we will look at what are the problems in CNN
Understand the current research trends in CNN
Materials to look at for CNN
RL Problems and Trends4:40
In this video, we will see what are the problems in RL
Understand the current research trends in RL
Materials to look at for RL
GAN Problems and Trends4:34
In this video, we will see what are the problems in GAN
Understand the current research trends in GAN
Materials to look at for GAN
Test your knowledge

Requirements

While knowledge of the Keras framework is not required, it is assumed that you’re well versed with the Machine Learning concepts and Python programming language.

Description

Keras is an Open source Neural Network library written in Python. It is a Deep Learning library for fast, efficient training of Deep Learning models. It is a minimal, highly modular framework that runs on both CPUs and GPUs and allows you to put your ideas into action in the shortest possible time. Because it is lightweight and very easy to use, Keras has gained quite a lot of popularity in a very short time.

This comprehensive 3-in-1 course takes a step-by-step practical approach to implement fast and efficient Deep Learning models: Projects on Image Processing and Reinforcement Learning. Initially, you’ll learn backpropagation, install and configure Keras to understand callbacks and customize the process. You’ll develop a deep learning network from scratch with Keras using Python to solve a practical problem of classifying the traffic signs on the road. Finally, you’ll get to grips with Keras to implement fast and efficient deep-learning models with ease.

Towards the end of this course, you'll use AI with Keras for building complex Deep Learning networks with fewer lines of coding in Python.

Contents and Overview

This training program includes 2 complete courses, carefully chosen to give you the most comprehensive training possible.

The first course, Deep Learning with Keras, covers implementing deep learning neural networks with Python. Keras is a high-level neural network library written in Python and runs on top of either Theano or TensorFlow. It is a minimal, highly modular framework that runs on both CPUs and GPUs and allows you to put your ideas into action in the shortest possible time. This course will help you get started with the basics of Keras, in a highly practical manner.

The second course, Hands-On Artificial Intelligence with Keras and Python, covers how to use AI with Keras for building complex Deep Learning networks with fewer lines of coding in Python. This course will help you learn by doing an industry relevant problem in image processing domain, develop and understand automation and AI techniques. You will learn how to harness the power of algorithms by creating apps which intelligently interact with the world around you, addressing common challenges faced in AI ecosystem. By the end of the course, you will be able to build real-world artificial intelligence applications using Keras and Python.

Towards the end of this course, you'll use AI with Keras for building complex Deep Learning networks with fewer lines of coding in Python.

About the Authors

Antonio Gulli is a software executive and business leader with a passion for establishing and managing global technological talent, innovation, and execution. He is an expert in search engines, online services, machine learning, information retrieval, analytics, and cloud computing. So far, he has been lucky enough to gain professional experience in four different countries in Europe and has managed people in six different countries in Europe and America. Antonio served as CEO, GM, CTO, VP, director, and site lead in multiple fields ranging from publishing (Elsevier) to consumer internet (Ask .com and Tiscali) and high-tech R&D (Microsoft and Google).
Sujit Pal is a technology research director at Elsevier Labs, working on building intelligent systems around research content and metadata. His primary interests are information retrieval, ontologies, natural language processing, machine learning, and distributed processing. He is currently working on image classification and similarity using deep learning models. Prior to this, he worked in the consumer healthcare industry, where he helped build ontology-backed semantic search, contextual advertising, and EMR data processing platforms. He writes about technology on his blog at Salmon Run.
Sandipan Das is working as a senior software engineer in the field of perception within the Autonomous vehicles industry in Sweden. He has more than 8 years of experience in developing and architecting various software components. He understands the industry needs and the gaps in between a traditional university degree and the job requirements in the industry. He has worked extensively on various neural network architectures and deployed them in real vehicles for various perception tasks in real-time.

Who this course is for:

Software Developers, Data Scientists with experience in Machine Learning or an AI Programmer with some exposure to Neural Networks: looking to achieve the power of Artificial Intelligence and want to build some broad range of skills such as image translation, autonomous driving simulation, Deep Reinforcement Learning with AI!

A Practical Guide to Deep Learning with Keras

What you'll learn

Explore related topics

Course content

Deep Learning with Keras9 lectures • 1hr 21min

Hands-On Artificial Intelligence with Keras and Python20 lectures • 2hr 31min

Requirements

Description

Who this course is for: