Hands on Deep Learning with Keras, Tensorflow, and Apache Spark™

Overview

This course is aimed at the practicing data scientist who is eager to get started with deep learning, as well as software engineers and technical managers interested in a thorough, hands-on overview of deep learning and its integration with Apache Spark.

The course covers the fundamentals of neural networks and how to build distributed TensorFlow models on top of Spark DataFrames. Throughout the class, you will use Keras, TensorFlow, Deep Learning Pipelines, and Horovod to build and tune models. This course is taught entirely in Python.

Each topic includes lecture content along with hands-on labs in the Databricks notebook environment.

Learning Objectives

After taking this class, students will be able to:

• Build a neural network with Keras
• Explain the difference between various activation functions and optimizers
• Track experiments with MLflow
• Apply models at scale with Deep Learning Pipelines
• Perform transfer learning
• Build distributed Tensorflow models with Horovod

Prerequisites

• Python (numpy and pandas)
  • Pandas Tutorial
  • Numpy Tutorial
• Apache Spark™ for Machine Learning and Data Science or equivalent experience

Delivery Requirements

• A computer or laptop
• Chrome or Firefox web browser – preferably Chrome
• Internet access with unfettered connections to the following domains:
  • *.databricks.com – required
  • keras.io – required
  • spark.apache.org – required

Topics

• Intro to Neural Networks with Keras
  • Neural network architectures
  • Activation functions
  • Evaluation metrics
  • Batch sizes, epochs, etc.
• MLflow
  • Reproducible ML/DL
• Convolutional Neural Networks
  • Convolutions
  • Batch Normalization
  • Max Pooling
  • ImageNet Architectures
• Deep Learning Pipelines
  • Model inference at scale
• Horovod
  • Distributed Tensorflow training
  • Ring-All Reduce

Audience

Data scientists, analysts, architects, software engineers, and technical managers who want to learn deep learning and apply it at scale using Apache Spark.