ELF OpenGo
20-0 against Go professionals, on a single GPU — then open-sourced
Latest: Final model + ICML 2019 analysis paper (v2, 2019)
FAIR's open reimplementation of AlphaZero for Go, built on the ELF reinforcement-learning platform. Trained on 2,000 GPUs over about two weeks, it went 20-0 against four top-30 professional players — while running on a single GPU — and released its code, trained models and analysis so anyone could reproduce a superhuman Go engine.
Why it matters
An open reimplementation of AlphaZero for Go that went 20-0 against four top-30 professionals while running on a single GPU, then released code, models, 20 million self-play games and a playable Windows binary so anyone could reproduce a superhuman engine. The repository was archived in December 2020.
Facts
- Beat pros 20-0 with 50 seconds per move while the humans had unlimited time; the team also released win-rate analyses of 87,000 professional human games — a gift to the Go community.
Try it yourself
Models, game analysis tool and Windows binary ↗ Archived code on GitHub ↗ Read the paper ↗
Sources
arXiv ↗GitHub · ELF ↗Meta AI ↗