Learning Ray

Index

Symbols

• @ray.remote decorator, Functions and remote Ray tasks, An Overview of the Ray Core API
• @serve.batch decorator, Request Batching
• @serve.deployment decorator, Model Serving, Defining a Basic HTTP Endpoint
  • tuning replicas and resource allocation for a deployment, Scaling and Resource Allocation

A

• abstractions
  • law of leaky abstractions, What Led to Ray?
  • provided by Ray, Simplicity and abstraction
• action distribution, Accessing policy and model states, Accessing policy and model states
• action space (RL), Recapping RL Terminology, Working with Multiple Agents
  • defining for policy server in RLlib environment, Defining a server
  • parametric action spaces in RLlib, Other Advanced Topics
• actions (RL), Recapping RL Terminology
  • computing in Python RLlib API, Computing actions
  • passed to steps in multi-agent environment, Working with Multiple Agents
  • probabilities of taking each action, Accessing policy and model states
  • simplifying assumptions about, Recapping RL Terminology
• actors, From classes to actors
  • controller actor for Ray Serve deployments, Architectural Overview
  • converting Simulation class to actor, Building a Distributed Ray App
  • GCS storing location of, The Head Node
  • mapping data with, support by Datasets, Computing Over Ray Datasets
  • Ray AIR usage combined with tasks for advanced composite workloads, Workloads That Are Suited for AIR
  • Ray Datasets, using in distributed batch inference, Distributed Batch Inference
  • in Ray Serve deployment, Architectural Overview
  • Ray patterns and anti-patterns for, Reducing ...