Book Image

Distributed Machine Learning with Python

By : Guanhua Wang
Book Image

Distributed Machine Learning with Python

By: Guanhua Wang

Overview of this book

Reducing time cost in machine learning leads to a shorter waiting time for model training and a faster model updating cycle. Distributed machine learning enables machine learning practitioners to shorten model training and inference time by orders of magnitude. With the help of this practical guide, you'll be able to put your Python development knowledge to work to get up and running with the implementation of distributed machine learning, including multi-node machine learning systems, in no time. You'll begin by exploring how distributed systems work in the machine learning area and how distributed machine learning is applied to state-of-the-art deep learning models. As you advance, you'll see how to use distributed systems to enhance machine learning model training and serving speed. You'll also get to grips with applying data parallel and model parallel approaches before optimizing the in-parallel model training and serving pipeline in local clusters or cloud environments. By the end of this book, you'll have gained the knowledge and skills needed to build and deploy an efficient data processing pipeline for machine learning model training and inference in a distributed manner.
Table of Contents (17 chapters)
1
Section 1 – Data Parallelism
6
Section 2 – Model Parallelism
11
Section 3 – Advanced Parallelism Paradigms

Wrapping up the whole model parallelism pipeline

In this section, we will discuss the components for implementing model parallelism. We will first discuss how to implement a model parallel training pipeline and then how to implement a model parallel serving pipeline.

A model parallel training overview

Let's look at a simple example of model parallel training, as shown in the following diagram:

Figure 7.1 – Model parallel training on a three-layer deep neural network (DNN) model

As shown in the preceding diagram, we have a three-layer DNN model, and we split each layer onto one GPU. For example, we put Layer 1 on GPU1 and Layer 2 on GPU2.

Forward propagation in model parallel training is shown on the left side of Figure 7.1. It works as follows:

  1. After GPU1 consumes the input training batch, it will calculate the activation values of Layer 1.
  2. After GPU2 receives output from GPU1, GPU2 starts its own forward propagation, which...