Skip to content

Implementation of the algorithm behind AlphaGo Zero to learn and play "Connect 4"

Notifications You must be signed in to change notification settings

philippwindischhofer/C4

Repository files navigation

Connect 4

This is an implementation of the algorithm behind AlphaGo Zero (Nature volume 550, pages 354–359, doi:10.1038/nature24270) to learn "Connect 4" purely from self-play and without any prior knowledge of the game.

To be used as follows:

Options:
  -h, --help            show this help message and exit
  -t, --train           train a new model or resume training of an already
                        existing model [if specified with -f]
  -f MODEL_FILE, --file=MODEL_FILE
                        specifies a model to use for training, benchmarking or
                        manual play
  -c, --computer        launches a new game against a model [specified with
                        -f]
  -m, --manual          launches a new game against a human opponent
  -b, --benchmark       benchmarks a model [specified with -f] against a
                        random player

If CUDA is installed and Tensorflow set up to run on a GPU, but you still want to force it to run on the CPU instead, simply run the main executable as

CUDA_VISIBLE_DEVICES="" optirun python ./main.py

example game

For slightly more background information, see this blog post.

About

Implementation of the algorithm behind AlphaGo Zero to learn and play "Connect 4"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages