BrainLearn chess engine

https://chess.massimilianogoi.com/download/brainlearn/

BrainLearn is a Stockfish-derived engine by amchess and like SugaR can learn and store the datas in a file called experience.bin. It can be considered an expansion of Brainfish with the ability to learn patterns from the games.

It's been some time I've taken a look at this engine, but I didn't hosted it before because the previous versions were buggy and not that strong... This time things were different: one of my tester (and the actual greatest contributor to this site), Constantin P., has made a tournament with a very impressive result:



therefore I've decided to host BrainLearn too.

The actual version is 5.2.1, but I can't host it on my website because its source has not been published, so it doesn't fulfill the GPL license requirements (the only exception I've made until now it's Raubfisch). I will include the BrainLearn source code in the downloadable package because it's not updated on its GitHub official page.

My executables are a little faster than the original ones:


This distribution is enhanced by the pgo-build option. BrainLearn automatically detects your maximum number of threads and set it as default value in the UCI settings.

BrainLearn comes with almost any architecture available, hence also for old 32bit operating systems like Windows XP, old machines, 64bit systems, standard machines, modern machines and power computers. It runs on Windows only this time.

Binaries of the latest version:

Windows:

- BrainLearn 5.2 X 32bit 32bit.exe FOR 32BIT WINDOWS SYSTEMS
- BrainLearn 5.2 X 32bit 32bit general.exe FOR 32BIT WINDOWS SYSTEMS
- BrainLearn 5.2 X 32bit 32bit old.exe FOR OLD MACHINES HAVING A 32BIT WINDOWS SYSTEM
- BrainLearn 5.2 X 32bit 32bit ppc-32.exe FOR POWER COMPUTERS HOSTING A WINDOWS SYSTEM 32BIT
- BrainLearn 5.2 X 32bit x64 bmi2.exe FOR 64BIT WINDOWS SYSTEM AND INTEL PROCESSORS
- BrainLearn 5.2 X 32bit x64 general.exe FOR 64BIT WINDOWS SYSTEM
- BrainLearn 5.2 X 32bit x64 modern.exe FOR POWERFUL PROCESSORS ON MACHINES HOSTING A 64BIT WINDOWS SYSTEM (POPULATION COUNT ALGORYTHM INCLUDED)
- BrainLearn 5.2 X 32bit x64 ppc-64.exe FOR POWER COMPUTERS HOSTING A WINDOWS SYSTEM 64BIT (POPULATION COUNT ALGORYTHM INCLUDED)


Last edited by jorge91 on Sun May 08, 2022 9:46 pm; edited 1 time in total

BrainLearn 5.2.1 X.

While traveling on train and at home the next days I followed amused and surprised the astonishing trend of views for this new engine, is it that good? It surpassed the sessions for my book page, that is quite odd for my website (apart the day of the Goi 6.2.1 release and the next one).

Since this engine is so requested I have decided to update it to the last executables although not being in possess of its last source. I will include the previous one in the package, anyway. I invite amchess to update the source on its GitHub page.

Any updates on here?

Unluckily not.

BrainLearn 6.1
For Windows only.

Changelog: BrainLearn 6.1, in particular: https://github.com/amchess/BrainLearn/commit/f1d4e2b42d2760860cb0c642328e997cec38b976.

This version is particularly strong.

Binaries of the latest version:

Windows:

- BrainLearn 6.1 32bit 32bit.exe FOR 32BIT WINDOWS SYSTEMS
- BrainLearn 6.1 32bit 32bit general.exe FOR 32BIT WINDOWS SYSTEMS
- BrainLearn 6.1 32bit 32bit old.exe FOR OLD MACHINES HAVING A 32BIT WINDOWS SYSTEM
- BrainLearn 6.1 32bit 32bit ppc-32.exe FOR POWER COMPUTERS HOSTING A WINDOWS SYSTEM 32BIT
- BrainLearn 6.1 32bit x64 bmi2.exe FOR 64BIT WINDOWS SYSTEM AND INTEL PROCESSORS
- BrainLearn 6.1 32bit x64 general.exe FOR 64BIT WINDOWS SYSTEM
- BrainLearn 6.1 32bit x64 modern.exe FOR POWERFUL PROCESSORS ON MACHINES HOSTING A 64BIT WINDOWS SYSTEM (POPULATION COUNT ALGORYTHM INCLUDED)
- BrainLearn 6.1 32bit x64 ppc-64.exe FOR POWER COMPUTERS HOSTING A WINDOWS SYSTEM 64BIT (POPULATION COUNT ALGORYTHM INCLUDED)

pay attention:
All these NN engines do not work on the Infinity chess gui properly.

Amigo disculpen: quiero descargar el BrainLearn 6.1 pero no lo logro. podrían ayudarme oh enviarme un enlace mas sencillo? LES VOY AGRADECER.

use this: https://github.com/amchess/BrainLearn/releases/tag/6.1
Windows, Linux compiles

I had some losses because of the engine bad calculation, even against Lc0. It didn't happen with Stockfish or others, I will remove it from the list of my site until the bug is corrected.

[pgn][Round "?"]
[White "Maxath, BrainLearn 6.1 64 B"]
[Black "Trasno, Lc0 v0.23.2"]
[Result "0-1"]
[ECO "C88"]
[WhiteElo "2434"]
[BlackElo "2512"]
[Annotator "0.06;?ev?"]
[PlyCount "128"]
[EventDate "2020.01.08"]
[SourceTitle "playchess.com"]
[TimeControl "300+1"]

{Lc0 v0.23.2 (4 cores): 17.5 plies; 107kN/s AMD Ryzen Threadripper 3970X
32-Core Processor 3900MHz, (32 cores, 64 threads), 1024 MB} 1. e4 {0.01/0 0} e5
{1} 2. Nf3 {0.01/0 0} Nc6 {1} 3. Bb5 {0.01/0 0} a6 {1} 4. Ba4 {0.01/0 0} Nf6 {1
} 5. O-O {0.01/0 0} Be7 {1} 6. Re1 {0.01/0 0} b5 {1} 7. Bb3 {0.01/0 0} O-O {1}
8. d3 {0.01/0 0} d6 {1} 9. a4 {0.01/0 0} b4 {1} 10. a5 {0.01/0 0} Be6 {1} 11.
Bxe6 {0.01/0 0} fxe6 {1} 12. h3 {0.01/0 0} Rb8 {1} 13. c3 {0.01/0 0} d5 {1} 14.
Nbd2 {0.01/0 0} Bc5 {1} 15. c4 {0.01/0 0} d4 {1} 16. Qa4 {0.01/0 0} Qe8 {1} 17.
Nb3 {0.01/0 0} Be7 {1} 18. Nh2 {0.01/0 0} h6 {1} 19. Bd2 {0.06/25 5} Nh5 {
1 (Qd7)} 20. g3 {0.08/29 16} Kh7 {1 (Rf7)} 21. Kg2 {0.10/28 8} Nf6 {8 (Qd7)}
22. Ng4 {0.06/31 19} Rd8 {10 (Rf7)} 23. Rh1 {0.11/28 19} Qd7 {9 (Qf7)} 24. Raf1
{0.17/25 4} Rb8 {5 (Qe8)} 25. Qa1 {0.14/27 8} Qe8 {1} 26. Rhg1 {0.11/29 16} Nd7
{13 (Qg6)} 27. f4 {0.20/25 4} h5 {17 (Rd8)} 28. Nh2 {0.19/27 6} h4 {1} 29. gxh4
{0.14/30 11} exf4 {1} 30. Bxf4 {0.14/29 6} Qh5 {1} 31. Ng4 {0.06/33 39} Bd6 {
1 (Rbc8)} 32. Bxd6 {0.18/28 16} cxd6 {17} 33. Qe1 {0.08/30 1} e5 {1 (Rxf1)} 34.
Qg3 {0.27/27 4} Ne7 {18 (Qg6)} 35. Nxd4 {0.00/28 14} exd4 {1} 36. Qxd6 {
0.00/30 3} Qe8 {1} 37. Rxf8 {0.00/31 3} Nxf8 {6} 38. Rf1 {0.00/35 4} Nc6 {5}
39. Rf6 {0.00/35 0} Kh8 {9 Maxath,BrainLearn 6.1 64 BM offers a draw} 40. Qxc6
{0.00/35 1} gxf6 {24} 41. Qxf6+ {0.00/36 0} Kg8 {6} 42. Qxd4 {0.00/34 5} Rd8 {
19} 43. Qb6 {0.00/38 0} Nh7 {13} 44. Qc7 {-0.07/30 0} Rd7 {4} 45. Qg3 {0.00/31
2} Rg7 {3} 46. Qd6 {0.00/35 3} Qf8 {22 (Qe7)} 47. Qxa6 {-0.33/31 44} Kh8 {5}
48. Qe6 {-0.45/29 27} Qf4 {1} 49. c5 {-0.48/29 5} Qd2+ {9 (Rg8)} 50. Kf3 {
-0.42/23 2} Qxd3+ {2} 51. Ne3 {-0.35/26 0} Nf8 {2} 52. Qd5 {-0.33/28 0} Rd7 {
2 (Qb1)} 53. Qxd3 {-0.39/26 4} Rxd3 {1} 54. e5 {-0.53/31 20} Ne6 {11} 55. c6 {
-0.64/34 0} b3 {3 (Rd2)} 56. Ke2 {-0.95/29 16} Rd4 {12} 57. Ng4 {-1.29/28 0}
Ra4 {9} 58. Kd2 {-1.64/28 1} Rxa5 {2} 59. Kc3 {-1.74/26 1} Rc5+ {2} 60. Kxb3 {
-1.89/29 8} Rxc6 {1} 61. h5 {-2.16/28 6} Kg7 {4 (Nd4+)} 62. h6+ {-32.54/31 1}
Kg6 {3} 63. h4 {-32.54/30 1} Rc8 {3 (Nd4+)} 64. Ka2 {-32.54/27 4} Kf5 {2 (Nd4)
Maxath,BrainLearn 6.1 64 BM resigns (Lag: Av=0.21s, max=0.7s)} 0-1[/pgn]

I have the same problems, but the authors don't want to hear that. On the Infinity platform, this engine doesn't work at all. Whether the problem is due to the gui or caused by the engine I can't say, because all other engines work fine. Also Marco Zerbinati doesn't want to accept this and prefers to ban his members if they make criticism. I don't use NN engines of any kind anymore, Sugar, Shashchess...

BrainLearn: Updated May, 25 2021.

For Windows and Linux, NO ANDROID, NO MACOS!

Changelog:

- MCTS evolved
- Concurrent Experience
- BrainLearnDGV (???)
- Stockfish patch with timestamp: 1621353983
- New NNUE architecture and net


This edition includes:


Windows:

- BrainLearn13-general-32.exe for 32-bit general Windows systems
- BrainLearn13-general-64.exe for 64-bit general Windows systems
- BrainLearn13-x86-32.exe for 32-bit Windows systems
- BrainLearn13-x86-64.exe for 64-bit Windows systems
- BrainLearn13-x86-64-avx2.exe for 64-bit Windows systems and AMD processors
- BrainLearn13-x86-64-avx512.exe for Knights Landing, Skylake-X and Cascade Lake processors
- BrainLearn13-x86-64-bmi2.exe for 64-bit Windows systems and Intel processors
- BrainLearn13-x86-64-modern.exe for modern machines on a 64bit Windows system (uses popcnt asm-instruction)
- BrainLearn13-x86-64-sse3-popcnt.exe for machines using popcnt asm-instructions on a 64-bit Windows system having CPU Opteron (since Stepping E4), Sempron (since Palermo. Stepping E3), Athlon 64 (since Venice Stepping E3 and San Diego Stepping E4), Athlon 64 FX (since San Diego Stepping E4), Athlon 64 X2, Phenom 64 X2, Turion family, K10 family, APU family (including without GPU), FX Series, Zen, Zen+, Zen2, Celeron D, Celeron (starting with Core microarchitecture), Pentium 4 (since Prescott), Pentium D, Pentium Extreme Edition (but NOT Pentium 4 Extreme Edition), Pentium Dual-Core, Pentium (starting with Core microarchitecture), Intel Core, Intel Xeon (since Nocona), Intel Atom, VIA/Centaur C7, VIA/Centaur Nano, VIA/CentaurTransmeta Efficeon TM88xx (NOT Model Numbers TM86xx)
- BrainLearn13-x86-64-sse41-popcnt.exe for machines using popcnt asm-instructions on a 64-bit Windows system having CPU Intel Core 2 or AMD Bulldozer-based FX CPUs
- BrainLearn13-x86-64-ssse3.exe for 64-bit Windows system having CPU Opteron (since Stepping E4), Sempron (since Palermo. Stepping E3), Athlon 64 (since Venice Stepping E3 and San Diego Stepping E4), Athlon 64 FX (since San Diego Stepping E4), Athlon 64 X2, Phenom 64 X2, Turion family, K10 family, APU family (including without GPU), FX Series, Zen, Zen+, Zen2, Celeron D, Celeron (starting with Core microarchitecture), Pentium 4 (since Prescott), Pentium D, Pentium Extreme Edition (but NOT Pentium 4 Extreme Edition), Pentium Dual-Core, Pentium (starting with Core microarchitecture), Intel Core, Intel Xeon (since Nocona), Intel Atom, VIA/Centaur C7, VIA/Centaur Nano, VIA/CentaurTransmeta Efficeon TM88xx (NOT Model Numbers TM86xx)



Linux:

- BrainLearn13-general-32 for 32-bit general Linux systems
- BrainLearn13-general-64 for 64-bit general Linux systems
- BrainLearn13-x86-32 for 32-bit Linux systems
- BrainLearn13-x86-64 for 64-bit Linux systems
- BrainLearn13-x86-64-avx2 for 64-bit Linux systems and AMD processors
- BrainLearn13-x86-64-avx512 for Knights Landing, Skylake-X and Cascade Lake processors
- BrainLearn13-x86-64-bmi2 for 64-bit Linux systems and Intel processors
- BrainLearn13-x86-64-modern for modern machines on a 64bit Linux system (uses popcnt asm-instruction)
- BrainLearn13-x86-64-sse3-popcnt for machines using popcnt asm-instructions on a 64-bit Linux system having CPU Opteron (since Stepping E4), Sempron (since Palermo. Stepping E3), Athlon 64 (since Venice Stepping E3 and San Diego Stepping E4), Athlon 64 FX (since San Diego Stepping E4), Athlon 64 X2, Phenom 64 X2, Turion family, K10 family, APU family (including without GPU), FX Series, Zen, Zen+, Zen2, Celeron D, Celeron (starting with Core microarchitecture), Pentium 4 (since Prescott), Pentium D, Pentium Extreme Edition (but NOT Pentium 4 Extreme Edition), Pentium Dual-Core, Pentium (starting with Core microarchitecture), Intel Core, Intel Xeon (since Nocona), Intel Atom, VIA/Centaur C7, VIA/Centaur Nano, VIA/CentaurTransmeta Efficeon TM88xx (NOT Model Numbers TM86xx)
- BrainLearn13-x86-64-sse41-popcnt for machines using popcnt asm-instructions on a 64-bit Linux system having CPU Intel Core 2 or AMD Bulldozer-based FX CPUs
- BrainLearn13-x86-64-ssse3 for 64-bit Linux system having CPU Opteron (since Stepping E4), Sempron (since Palermo. Stepping E3), Athlon 64 (since Venice Stepping E3 and San Diego Stepping E4), Athlon 64 FX (since San Diego Stepping E4), Athlon 64 X2, Phenom 64 X2, Turion family, K10 family, APU family (including without GPU), FX Series, Zen, Zen+, Zen2, Celeron D, Celeron (starting with Core microarchitecture), Pentium 4 (since Prescott), Pentium D, Pentium Extreme Edition (but NOT Pentium 4 Extreme Edition), Pentium Dual-Core, Pentium (starting with Core microarchitecture), Intel Core, Intel Xeon (since Nocona), Intel Atom, VIA/Centaur C7, VIA/Centaur Nano, VIA/CentaurTransmeta Efficeon TM88xx (NOT Model Numbers TM86xx)

Hopely there would be an Android version for this engine.. Amuen

Zippo wrote:

Hopely there would be an Android version for this engine.. Amuen  

No: Android doesn't allow to create/write an experience file, as far as I know. Something about file permissions.

Owhh okay.. thx for the confirm Sir..

csnzax wrote:

https://www.dropbox.com/s/mjanpnkubf8dwm2/BrainLearn-12.1-armv7-neon?dl=0

Thank's for the engine dear..
But do u have the 64bit version ? ( arm8 )
Sinced my smarrphone waa running in 64 bit

Thank's in advance

No way it will work properly.

csnzax wrote:

https://www.dropbox.com/s/f6z21qt5ws6816j/BrainLearn-12.1-aarch64.7z?dl=0&file_subpath=%2FBrainLearn-12.1-aarch64

The link doesn't containt any material to download buddy.. it shows that the file can't be previewed.. but no option to download at all..

Download button is at the top right corner.

csnzax wrote:

Download button is at the top right corner.

The download does not appear, and there is no desire to download the application or enter.

Now I managed to download, but the network is old.

Mirror:

BrainLearn-12.1-aarch64

https://pixeldrain.com/u/o6hjxrCG

PS: Use pixeldrain, it's the most convenient and easiest choice.