chessforyou Bettina&Terry77
chessforyou Bettina&Terry77
chessforyou Bettina&Terry77
Would you like to react to this message? Create an account in a few clicks or log in to continue.

chessforyou Bettina&Terry77


 
HomeLatest imagesRegisterLog in
WELCOME TO FORUM OF Angels77 * named in memory of Bettina & Terry
Search
 
 

Display results as :
 
Rechercher Advanced Search
Search
 
 

Display results as :
 
Rechercher Advanced Search
Latest topics
» Latest abrok Development Version March 22nd (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 24, 2024 7:05 pm by LondonFrau

» 1337 pro ctg update
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyThu Mar 21, 2024 8:44 am by Guest

» Latest abrok Development Version March 20th (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyWed Mar 20, 2024 5:45 pm by LondonFrau

» 26 plus elo increase ..Latest SF from Abrok March 14th The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 17, 2024 1:41 pm by LondonFrau

» Latest SF from Abrok March 7th
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 10, 2024 5:52 pm by LondonFrau

» Stockfish 16.1
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 03, 2024 2:33 am by LondonFrau

»  Update default main net to nn-b1a57edbea57.nnue Febuary 17th (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyTue Feb 20, 2024 2:41 am by LondonFrau

» Latest abrok Development Version Febuary 3rd (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyWed Feb 07, 2024 9:19 am by LondonFrau

»  abrok Development Version January 26th (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySat Jan 27, 2024 1:55 pm by LondonFrau

Latest topics
» Latest abrok Development Version March 22nd (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 24, 2024 7:05 pm by LondonFrau

» 1337 pro ctg update
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyThu Mar 21, 2024 8:44 am by Guest

» Latest abrok Development Version March 20th (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyWed Mar 20, 2024 5:45 pm by LondonFrau

» 26 plus elo increase ..Latest SF from Abrok March 14th The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 17, 2024 1:41 pm by LondonFrau

» Latest SF from Abrok March 7th
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 10, 2024 5:52 pm by LondonFrau

» Stockfish 16.1
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySun Mar 03, 2024 2:33 am by LondonFrau

»  Update default main net to nn-b1a57edbea57.nnue Febuary 17th (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyTue Feb 20, 2024 2:41 am by LondonFrau

» Latest abrok Development Version Febuary 3rd (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyWed Feb 07, 2024 9:19 am by LondonFrau

»  abrok Development Version January 26th (SF The gift that keeps on giving
MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptySat Jan 27, 2024 1:55 pm by LondonFrau

Navigation
 Portal
 Index
 Memberlist
 Profile
 FAQ
 Search
Navigation
 Portal
 Index
 Memberlist
 Profile
 FAQ
 Search
Forum
Forum
Affiliates
free forum
 


Affiliates
free forum
 



 

 MoonFish-NN 111119 32-64 Jörg Oster, Marco Zerbinati

Go down 
AuthorMessage
Doppelganger
Dark Angels
Doppelganger


Male Posts : 562
Reputation : 2159
Join date : 2010-02-03
Location : We often take for granted the very things that most deserve our gratitude. Cynthia Ozick

MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati Empty
PostSubject: MoonFish-NN 111119 32-64 Jörg Oster, Marco Zerbinati   MoonFish-NN 111119 32-64    Jörg Oster, Marco Zerbinati EmptyMon Nov 11, 2019 11:59 pm

Add Self-Learning

  • Save hash Capability
  • BIN Book Support
    -Large Pages Support


 Assets 6

[You must be registered and logged in to see this link.] 1.29 MB
[You must be registered and logged in to see this link.] 1.13 MB
[You must be registered and logged in to see this link.] 1.14 MB
[You must be registered and logged in to see this link.] 1.13 MB







[You must be registered and logged in to see this link.]Overview


MoonFish-NN is a free UCI chess engine derived from Moonfish original code by Jörg Oster [[You must be registered and logged in to see this link.]]. It is not a complete chess program and requires some UCI-compatible GUI (e.g. XBoard with PolyGlot, eboard, Arena, Sigma Chess, Shredder, Chess Partner, Aquarium or Fritz) in order to be used comfortably. Read the documentation for your GUI of choice for information about how to use MoonFish-NN with it.

This version of MoonFish-NN supports up to 128 cores. The engine defaults to one search thread, so it is therefore recommended to inspect the value of the Threads UCI parameter, and to make sure it equals the number of CPU cores on your computer.

This version of MoonFish-NN has support for Syzygybases.

[You must be registered and logged in to see this link.]Files


This distribution of MoonFish-NN consists of the following files:


  • Readme.md, the file you are currently reading.
  • Copying.txt, a text file containing the GNU General Public License.
  • source, a subdirectory containing the full source code, including a Makefile that can be used to compile MoonFish-NN on Unix-like systems.


[You must be registered and logged in to see this link.]Uci options


[You must be registered and logged in to see this link.]Skill Level


Lower the Skill Level in order to make Stockfish play weaker (see also UCI_LimitStrength). Internally, MultiPV is enabled, and with a certain probability depending on the Skill Level a weaker move will be played.



Enable weaker play aiming for an Elo rating as set by UCI_Elo. This option overrides Skill Level.



If enabled by UCI_LimitStrength, aim for an engine strength of the given Elo. This Elo rating has been calibrated at a time control of 60s+0.6s and anchored to CCRL 40/4.

[You must be registered and logged in to see this link.]ICCF Analyzes


Default: 0, Min: 0, Max:8

0, no MultiPV.

For analysis (purpose)


  • Value 1 corresponds to multiPV = 2
  • Value 2 to multiPV = 4
  • Value 3 to multiPV = 8
  • Value 4 to multiPV =16
  • Value 5 to multiPV = 32
  • Value 6 to multiPV = 64
  • Value 7 to multiPV = 128
  • Value 8 to multiPV = 256


1-8 MultiPV: higher depths and longer time to reach them. So, fewer tactical shots missed, but loss of some ELO, increasingly until 8, corresponding to multiPV = 256.

Recommended values: from 2 to 5 ( > 5 too wide search width)

[You must be registered and logged in to see this link.]MoonFish-NN can use two parallel books


original code by Thomas Zipproth: [You must be registered and logged in to see this link.]

[You must be registered and logged in to see this link.]NN section (Experimental Neural Networks inspired technics)


Experimental, MonteCarloTreeSearch, if activated, the engine's behaviour is similar to AlphaZero concepts. Idea are implemented, integrated on MoonFish-NN:

[You must be registered and logged in to see this link.]NN Persisted Self-Learning


Boolean, Default: True


  • [[You must be registered and logged in to see this link.]] MoonFish-NN implements a persistent learning algorithm, managing a file named experience.bin by Kelly kyniama and Andrea Manzo.


It is a collection of one or more positions stored with the following format (similar to in memory Stockfish Transposition Table):


  • best move
  • board signature (hash key)
  • best move depth
  • best move score


This file is loaded in an hashtable at the engine load and updated each time the engine receive quit or stop uci command. When SugaR-NN starts a new game or when we have max 8 pieces on the chessboard, the learning is activated and the hash table updated each time the engine has a best score at a depth >= 4 PLIES, according to Stockfish aspiration window.

At the engine loading, there is an automatic merge to experience.bin files, if we put the other ones, based on the following convention:

.bin

where


  • fileType="experience"/"bin"
  • qualityIndex , an integer, incrementally from 0 on based on the file's quality assigned by the user (0 best quality and so on)


N.B.

Because of disk access, to be effective, the learning must be made at no bullet time controls (less than 5 minutes/game).

[You must be registered and logged in to see this link.]NN Perceptron Algorithm


Boolean, Default: False


  • [[You must be registered and logged in to see this link.]] ( Perceptron Sigmoid activation by Stefano Cardanobile) for Late Move Reductions search as training signal


[You must be registered and logged in to see this link.]NN MCTS Score


Boolean, Default: False


  • [[You must be registered and logged in to see this link.]] ( (Montecarlo Tree Search Scores) by Jörg Oster) in main search function to an upper node.




[You must be registered and logged in to see this link.]
Back to top Go down
 
MoonFish-NN 111119 32-64 Jörg Oster, Marco Zerbinati
Back to top 
Page 1 of 1

Permissions in this forum:You cannot reply to topics in this forum
chessforyou Bettina&Terry77 :: ENGlNES :: ALL ENGlNES :: ENGlNES & CODES :: Stockfish :: StockFish Forks-
Jump to: