Skip to content

MuKoe

An open MuZero implementation orchestrated with Ray

Visit official site

What MuKoe is

The framework is described as a fully open-source MuZero implementation that uses Ray as a distributed orchestrator on GKE.

What you can do with it

  • Run distributed computing tasks
  • Accelerate learning with TPU hardware
  • Perform CPU inference and batch TPU inference

Requirements

  • CPU and TPU execution are both described
  • Ray manages the cluster on GKE
  • Specific hardware requirements are noted

Pros & cons

✓ What we like

  • The codebase is fully open source
  • An Atari game example is provided

! What to watch out for

  • Specific hardware requirements are stated
  • Orchestration is described on GKE with Ray

FAQ

What does it run on?

The material names CPU and TPU, with specific hardware requirements.

Where is the overview given?

A Google Next 2024 talk is described as giving a codebase overview.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review