The use of distributed programming is increasing. High availability requires multiple machines and often multiple data centers. Machine learning and AI models are run as parallel tasks on clusters to reduce training time. But distributed programming has always been hard to do--at least, until now. This report shows you an easier way.
Dean Wampler from Anyscale introduces you to Ray, an open source project that provides a concise and intuitive Python API for defining tasks that need to be distributed. Built by researchers at UC Berkeley, Ray does most of the tedious work of running workloads at massive scale. For the majority of distributed workloads, this guide shows you how Ray provides a flexible, efficient, and intuitive way to get work done.
- Learn how Ray builds on familiar language concepts: functions for stateless tasks and classes for stateful computing
- Use Ray libraries for reinforcement learning, hyperparameter tuning, distributed training of TensorFlow and PyTorch models, and model serving
- Work with Ray for general application development, including conventional microservices and serverless applications
- Get hands-on instruction using live Ray code with Dean Wampler's Meet the Expert session on O'Reilly online learning
Table of contents
- 1. Distributed Computing Is Hard but Necessary
- 2. The Ray API
- 3. Machine Learning Libraries That Use Ray
- 4. Ray for Applications
- 5. Recap and Next Steps
- Title: What Is Ray?
- Release date: September 2020
- Publisher(s): O'Reilly Media, Inc.
- ISBN: 9781492085751
You might also like
40 Algorithms Every Programmer Should Know
Learn algorithms for solving classic computer science problems with this concise guide covering everything from fundamental …
Head First Design Patterns, 2nd Edition
You know you don’t want to reinvent the wheel, so you look to design patterns—the lessons …
51+ hours of video instruction. Overview The professional programmer’s Deitel® video guide to Python development with …
High Performance Python, 2nd Edition
Your Python code may run correctly, but you need it to run faster. Updated for Python …