Skip to content

A neural evaluation function for chess engines and a companion trainer

Notifications You must be signed in to change notification settings

puneethnarayana/rStock

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

rStock

Stockfish is an open source UCI chess engine, developed by Tord Romstad, Joona Kiiski and Marco Costalba and licensed under the GNU General Public License version 3. The current version 4 (as of August 20, 2013) is available as C++ source code, and also has precompiled versions for Microsoft Windows, Mac OS X, and Linux 32-bit/64-bit. It can use up to thirty-two CPU cores in multiprocessor systems. The maximum size of the transposition tables is eight gigabytes. Stockfish implements an advanced alpha-beta search and uses bitboards.

This project modifies Stockfish and plugs in a neural evaluation function and a companion trainer for purely academic purposes. I thank Andrey Kotlarski without whose BPN, this project would be a lot harder to implement.

TODO:

  1. Convert the code and run it in Python
  2. Write a blog about the experience

About

A neural evaluation function for chess engines and a companion trainer

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published