jonathan-laurent / AlphaZero.jl

A generic, simple and fast implementation of Deepmind's AlphaZero algorithm.

Home Page:https://jonathan-laurent.github.io/AlphaZero.jl/stable/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Success stories

StepHaze opened this issue · comments

Can anyone share his (or her) success story related to AlphaZero.jl ?

I trained Connect4 bot, and it won all games against Connect4 mobile apps that I found on Google Play.

Of course, you are always welcome to share user stories (whether positive or negative)!

I made a little strategy game called Rektang. The solo mode is powered by several instances of an AlphaZero.jl bot which I trained up to 50 iterations. I think it is quite good now.
I used the AlphaZero.jl and Genie.jl packages to build an API server, which uses these instances to serve moves to any live game running in an HTML5 app at https://rektang.com.
To try the AI, just select "Solo Game" and start a game. Optionally, you can connect (there are limited options for the moment) in order to track your results as a Elo score.
NB. There is a short tutorial to learn the very simple rules.
Enjoy! Feedback welcome.
Huge THANK YOU to Jonathan for his help when I requested it, and to everyone involved in this AMAZING framework. 😘