About This Course

It’s no secret that deep learning has taken the industry by storm. With the ability to solve complex problems across wide project domains, this machine learning technique is here to stay. Whether attempting to produce a human-like translation of text or trying to solve a 50-year-old protein-folding problem, one approach has been at the forefront of AI advancements and serves as the foundation of all deep learning: the neural network.

Enthought’s Deep Learning course will cover the basic components that comprise neural networks and provide students with hands-on experience employing this popular machine learning technique. Students will walk away with a practical knowledge of how to implement neural networks in the context of their work. 

Over the course of this training, students will build an actionable understanding of deep learning concepts that go beyond neural networks and have implications in the larger machine learning ecosystem. By the end, scientists and engineers who participated will have the skills and confidence needed to navigate the world of deep learning.

This course is instructor-led. Consult the class schedule below for times and locations.

Register Now



Boost Your Career
Return on Investment
Practical and Applied
Deep learning is one of the most sought-after fields in machine learning and artificial intelligence today. Leveling up in this field will not only give you more tools for solving complex problems, it will also boost your career. Learn how to solve more complex problems, and make your workflows more efficient and productive, from Instructors with years of Industry experience applying deep learning techniques to science and engineering workflows. Create, train, and evaluate models, following best practices. Learn how to do it yourself, and fix it yourself. Walk away from this course with foundational mental models, not just how some API works.



Day 1: Introduction to Neural Networks

This course will begin with a survey of deep learning tools, methods, and techniques. Day one will focus on building an intuitive understanding of what deep learning is and how it relates to the larger field of machine learning. Students will develop a practical knowledge of neural networks and gain hands-on experience employing such models with Tensorflow and Keras. Day one will close with an exercise where students are given the opportunity to build their first neural network model from beginning to end.

Day 2: Training Neural Network Models

The second day will focus on a specific piece of the deep learning pipeline: the training loop. Here, students will explore advanced topics like utilizing TensorFlow’s GradientTape API to conduct automatic differentiation and how to create custom callbacks for monitoring internal states of models during training. Day two will close with a demonstration of how to develop a specialized model class with a custom training loop for creating a Generative Adversarial Network (GAN).

Day 3: Evaluating Neural Network Models

The course will close with a machine learning topic that is often overshadowed by the model building phase: how to accurately evaluate a model’s performance. Students will be taught the common metrics used for both classification and regression problems. Through the use of hand-on exercises, students will learn the common issues that plague neural networks as well as some practical approaches to remedy poor model performance.


Class Schedule

If you registered to attend this course online, the session times will be sent to you one week before your program start date. Virtual classes will be held on GoToMeeting.

Onsite corporate classes are also available. Discounts are available for 3 or more attendees and academics currently at a degree-granting institution. Contact us using the form on this page to learn more.

Where When Price (per person) Reserve a Seat
Online - Live Virtual March 14-16 | 9:00-11:00AM and 1:00-3:00PM MDT $1,250 Register Online
Online - Live Virtual June 27-29 | 9:00-11:00AM and 1:00-3:00PM EDT $1,250 Register Online

Contact Us

Questions or need help registering? Call us at 512.536.1057 or fill out the form:

    Course Syllabus & Topics

    This class is taught in real-time by an Enthought Trainer.



    Course Prerequisites

    What You’ll Learn

    Collapse All
    • Introduction to neural networks
      • Exploring end to end implementation of model
      • Defining forward propagation and backpropagation
      • Using TensorFlow and Keras API
      • Highlighting model components (layers and activation functions)
    • Building your first neural network
      • Sequential model building
      • Compiling and fitting a model
      • Capturing and analyzing model training history
    • Tuning and saving models
      • Diagnosing performance
      • Common remedies for neural network underperformance
      • Saving neural networks and loading pre-trained models
    • Mechanics of fitting a neural network model
      • Review loss function and optimization techniques
      • Introduce the neural network update step
      • Illustrate automatic differentiation with TensorFlow’s GradientTape
    • Preprocessing data for model training
      • Discuss loading and preparing data sets
      • Explore splitting data into training, validation, and testing subsets
      • Introduce data epochs and batch sizes, and batch training
    • Customizing training processes
      • Explore training loops with manual reporting
      • Demonstrate adding additional metrics to existing training routines
      • Leverage existing Keras callbacks and create custom callbacks
      • Illustrate complex training loop for generative adversarial network
    • Utilizing a neural network’s training history
      • Capturing and analyzing model training history
      • Diagnosing common patterns (overfitting, underfitting, early convergence)
    • Remedying common pitfalls in neural network underperformance
      • Explore quick fixes to improve model performance
      • Demonstrate advanced approaches to increase model performance
    • Evaluating final state of a neural network model
      • Define characteristics of a good (or bad) model
      • Explore various evaluation metrics and appropriate applications of each