ELF OpenGo

ELF OpenGo is a reimplementation of AlphaGoZero / AlphaZero that was trained on 2,000 GPUs over a 15 day period, and has achieved super-human performance. With only a single GPU, the ELF OpenGo bot played with four top 30 professional players and won 20-0 in slow games that impose no constraints on time spent for human players.

• Read our latest blog post on the Facebook AI Blog.

• Learn more about the initial release of ELF OpenGo on our Facebook Research Blog.

Human Professional Game Analysis

Using ELF OpenGo, we analyzed historical professional games from the past few centuries, and developed interactive tools to visualize the results.

• Interactive Professional Game Analysis

• Analysis of professional games (SGFs, CSVs)

Play a game with ELF OpenGo

ELF OpenGo public binary, allowing anyone with a CUDA-enabled GPU to play against the final ELF OpenGo model.

ELF OpenGo Windows binary

Code & Pre-Trained Models

We have open sourced the code used to train ELF OpenGo and released the models and data.

• Code for ELF OpenGo

• Pre-trained models:

  • Final model that is trained from scratch in 15 days using 2000 GPUs

  • Prototype model that won 20:0 against top professional players

  • 20 million self-play games, obtained from final model training

  • Analysis of self-play games (SGFs, CSVs)

Research Paper

• OpenGo: An Analysis and Open Reimplementation of AlphaZero, Yuandong Tian, Jerry Ma, Qucheng Gong, Shubho Sengupta, Zhuoyuan Chen, James Pinkerton, C. Lawrence Zitnick, arXiv 2019.