ray
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
What is ray?
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
Overview of ray
What it is
Ray is a unified framework for scaling AI and Python applications. It consists of a core distributed runtime and a set of AI libraries designed to simplify machine learning compute. Ray runs on any machine, cluster, cloud provider, and Kubernetes.
Key features
- Core distributed runtime with key abstractions: Tasks (stateless functions), Actors (stateful worker processes), and Objects (immutable values accessible across the cluster).
- AI Libraries: Ray Data (scalable datasets for ML), Ray Train (distributed training), Ray Tune (scalable hyperparameter tuning), RLlib (scalable reinforcement learning), and Ray Serve (scalable and programmable serving).
- Monitoring and debugging tools: Ray Dashboard for monitoring apps and clusters, and Ray Distributed Debugger for debugging Ray applications.
- Seamless scaling: the same Python code can run on a laptop or a cluster without modification.
- General-purpose: can performantly run any kind of workload written in Python.
Use cases
- Distributed training of machine learning models using Ray Train.
- Hyperparameter optimization at scale with Ray Tune.
- Serving machine learning models with Ray Serve.
- Reinforcement learning experiments using RLlib.
- Scaling general Python applications from a single node to a cluster.
Getting started
Install Ray with pip: pip install ray For nightly wheels, see the Installation page in the repository.
License and pricing
Ray is open source under the Apache-2.0 license. There is no pricing; it is free to use.
Based on the project's public README and repository metadata.