stockfish of the neural network

stockfish of the neural network

https://dropmefiles.com/VxrkL

stockfish of the neural network

If someone wants to try Stockfish with the nnue neural network, which was built in instead of the usual evaluation function, then I have collected a special archive. It contains only the most necessary and up-to-date files. Download and unpack at the link below. For AMD Ryzen processors, you need to run the avx2 file, and for not too old Intel processors, the bmi2 file. For other Protsiv alas, nothing yet. But even for these processors, it is not a fact that it will work.

The level of the game in the area of the usual Stockfish 10-11. Maybe even beyond these borders-the results of testers vary greatly. It is possible that they may depend on a specific processor model. The speed of the new version of the drain is almost half that of the usual One. For engine matches, we recommend setting the Slow Mover parameter to 42 or 60 in the settings.

Try and check:

https://dropmefiles.com/VxrkL[/quote]

from dear Rom 77[/quote]

create the eval folder and move the network there

https://cdn.discordapp.com/attachments/724700045525647420/728697754154565642/0307M_t1.7z

Yes, and the network needs to be renamed nn.bin, you need to run the isready command through the console to load the network.

SF-NNUE 07-07-2020

recycle network load
(loads the network via ucinewgame () instead of isready ())

"Nn.bin" is the default, but no longer hard-coded,
that is, now it can accept any file name through the UCI "EvalFile" option
(view file system)

the archive contains much faster popcnt executables (with sse4.1 support)

https://github.com/FireFather/sf-nnue/releases

sf-nnue 07-08-2020

updated binaries
fixes learn module

https://github.com/FireFather/sf-nnue/releases

https://dropmefiles.com/OL9Y5

Network for Stockfish

Gekkehenker 2020-06-27 - this is the latest current network at the moment

https://dropmefiles.com/kMcWJ

@lisandr inbox ples

download and install it as a normal engine,and everything is ready to use.Network inside .This is a very powerful engine with a network. enter the isready command in the console. It must be so



https://workupload.com/file/ggEUrvNVgmH

Thanks !

Ready to use+Network inside ......   This is the first time I've had the nn.bin load !

---------------

Android version SF_nnue_armv8_api24

https://dropmefiles.com/d9pdy

Download the network, rename to nn.bin. And specify the full path in the engine parameter. If everything is done correctly, then the score in the starting position will not be zero.

https://dropmefiles.com/Nm4MP - Video instruction

https://www.comp.nus.edu.sg/~sergio-v/nnue/ - Here are the nets

stockfish-NNUE-flto-avx512.exe- https://cdn.discordapp.com/attachments/724700045525647420/740257384290582588/stockfish-NNUE-flto-avx512.exe


stockfish-NNUE-flto-bmi2.exe- https://cdn.discordapp.com/attachments/724700045525647420/740257415534215260/stockfish-NNUE-flto-bmi2.exe

https://abrok.eu/stockfish/


That's it. There is no more division into Stockfish_Classical and Stockfish_NUE - there is one Stockfish.

New Neural Net - 20-09-03 12:15:46

https://tests.stockfishchess.org/api/nn/nn-308d71810dff.nnue

Processor:Intel Core i5-7600K, GHz:4.20 , Сore:1, Gui:Fritz-11, Games:100, Time:10sec+1sec, Book:Off, Endgame tables:Syzygy 3-4-5, Hash:128 MB, Ponder:Off.

Games-

https://pixeldrain.com/u/5s6NvftK

seems they are all same power

I've been doing these tests for a long time. Test the first networks and the last ones yourself and see the difference.

Author: SFisGOD
Date: Wed Oct 28 08:13:34 2020 +0100
Timestamp: 1603869214

Update default net to nn-2eb2e0707c2b.nnue

Optimization of the net weights of the 32 x 32 layer (1024 parameters) and net biases of the 512 x 32 layer (32 parameters) using SPSA.

Tuning of 32 x 32 Layer (800,000 games, 5 seconds time control):
https://tests.stockfishchess.org/tests/view/5f942040d3978d7e86f1aa05

Tuning of 512 x 32 Layer (80,000 games, 20 seconds time control):
https://tests.stockfishchess.org/tests/view/5f8f926d2c92c7fe3a8c608b

STC:
LLR: 2.96 (-2.94,2.94) {-0.25,1.25}
Total: 17336 W: 1918 L: 1754 D: 13664 Elo +3.29
Ptnml(0-2): 79, 1344, 5672, 1480, 93
https://tests.stockfishchess.org/tests/view/5f9882346a2c112b60691b34

LTC:
LLR: 2.94 (-2.94,2.94) {0.25,1.25}
Total: 37304 W: 1822 L: 1651 D: 33831 Elo +1.59
Ptnml(0-2): 27, 1461, 15501, 1640, 23
https://tests.stockfishchess.org/tests/view/5f98a4b36a2c112b60691b40

closes https://github.com/official-stockfish/Stockfish/pull/3201

Bench: 3403528

https://abrok.eu/stockfish/

You can download it separately from here: nn-2eb2e0707c2b.nnue

https://tests.stockfishchess.org/nns

Author: mstembera
Date: Sun Nov 1 08:02:40 2020 +0100
Timestamp: 1604214160

Update default net to nn-cb26f10b1fd9.nnue

Result of https://tests.stockfishchess.org/tests/view/5f9a06796a2c112b60691c0f tuning.

STC
LLR: 2.94 (-2.94,2.94) {-0.25,1.25}
Total: 53712 W: 5776 L: 5561 D: 42375 Elo +1.39
Ptnml(0-2): 253, 4282, 17604, 4431, 286
https://tests.stockfishchess.org/tests/view/5f9c7bbc6a2c112b60691d4d

LTC
LLR: 2.97 (-2.94,2.94) {0.25,1.25}
Total: 80184 W: 4007 L: 3739 D: 72438 Elo +1.16
Ptnml(0-2): 58, 3302, 33130, 3518, 84
https://tests.stockfishchess.org/tests/view/5f9d01f06a2c112b60691d87

closes https://github.com/official-stockfish/Stockfish/pull/3209

bench: 3517795

https://abrok.eu/stockfish/

Which is best engine for android for blitz games ??