TensorFlow and Keras Training Overview
Accelebrate's Machine Learning with TensorFlow & Keras training teaches attendees how to develop real-world machine learning applications powered by TensorFlow and Keras, popular technologies for building production-grade deep learning models. Students also learn how to use an HTTP API to retrieve model predictions.
Location and Pricing
Accelebrate offers instructor-led enterprise training for groups of 3 or more online or at your site. Most Accelebrate classes can be flexibly scheduled for your group, including delivery in half-day segments across a week or set of weeks. To receive a customized proposal and price quote for private corporate training on-site or online, please contact us.
In addition, some courses are available as live, instructor-led training from one of our partners.
Objectives
- Understand the necessary technical requirements for developing a TensorFlow-powered application
- Gain knowledge on different deep learning architectures and start to develop an intuition on what to chose for your problem
- Experiment with different deep learning algorithms
- Train and evaluate the performance of deep learning models
- Combine your model with Flask to create an HTTP API that returns model predictions
Prerequisites
All students should have programming experience, familiarity with Python 3, and basic knowledge of web-applications (i.e., how the HTTP protocol works). Knowledge of the following topics are useful, but not required:
- Machine learning concepts such as the bias and variance tradeoff, and accuracy estimation methods
- Familiarity with other machine learning algorithms such as Support Vector Machines, Adaptive Boosting (a.k.a. "AdaBoost"), Naïve Bayes, and Logistic Regression
- Knowledge of statistics, probability, linear algebra, and calculus
- Experience with data-provisioning systems, including file systems (local and remote) and databases (SQL and NoSQL)
Outline
Expand All | Collapse All
Introduction
- Overview neural networks and deep learning
- Organizing a project directory
- Setting-up programming environment
- Installing Python 3
- Creating a Python virtual environment
- Installing application dependencies
- Configuring TensorFlow, Keras, and TensorBoard
Model Architecture
- Load and prepare Bitcoin dataset
- Load CSV file into memory
- Extracting and exploring dataset features
- Split dataset into a train and a test set
- Overview of a TensorFlow computation graph (i.e. a "model")
- Overview of the Keras interface
- Instantiating a model
- Adding components
- Creating a queue and parallelize computations
- Using the debugger
- Deploying TensorBoard
Model Training and Evaluation
- cl Training a model
- Epochs
- Weights and Biases
- Hidden layers
- Learning rate and optimization (i.e. activation functions)
- Testing a model
- Measuring accuracy and error rates
- Determining when to stop
- Model deployment
Productization
- Overview of an (extremely) simple Flask application
- Creating an HTTP endpoint
- Retrieving JSON results
Conclusion
Training Materials
All TensorFlow & Keras training students receive comprehensive courseware covering all topics in the class.
Software Requirements
- Windows, Mac, or Linux
- Anaconda or comparable Python distribution with TensorFlow and Keras installed (Accelebrate will provide instructions)
- PyCharm, Spyder, or comparable Python IDE
- Related lab files that Accelebrate will provide
