| Summary: | Advances in machine learning techniques, the cloud, and the ability to leverage hardware acceleration have changed the way we work with data - adding entirely new capabilities and business models to the mix. But the demand for processing training data has outpaced the increase in computation power. This practical and comprehensive guide will show you how to distribute your machine learning workload across multiple machines and turn centralized systems into distributed ones. Machine Learning with Spark examines various technologies for building end-to-end distributed machine learning platforms based on the Apache Spark ecosystem with Spark MLlib, TensorFlow, Horovod, PyTorch, and more. This book shows you when to use each technology and why. You'll also learn how to: Build efficient parallelization of the training process Create a coherent model Leverage a set of open source tools to build scalable end-to-end ML platform Enable more advanced, tailor-made products Use distributed ML techniques to increase the quality of predictions and ML modules Design practical distributed machine learning systems
|
|---|