Stockfish 250823 Self Learning. Update until 24 August 2023 "Play turbulent when defending, simpler when attacking". Features: Self Learning Eman style as Stockfish.exp (can be renamed), CTG/Polyglot Books, MinimumThinkingTime. External network.
Download Windows 64-Bit (AVX2, BMI2, Modern, SSE41, source code):
https://solistachess.jimdosite.com/solista-news/
Compile with network dimension 2560 "nn-1ceb1ade2048.nnue":
https://solistachess.jimdosite.com/solista-news/
Have a nice day.
Eduard
Stockfish dev sl
-
Solista
- Posts: 202
- Joined: Thu Jul 06, 2023 11:47 am
- Location: Germany
- Has thanked: 7 times
- Been thanked: 293 times
- Contact:
Re: Stockfish dev sl
Stockfish 250823 Self Learning, direct download links:
Dimension 2048 (default):
https://pixeldrain.com/u/bScE8ANN
Dimension 2560:
https://pixeldrain.com/u/FoEbKSTL
Dimension 2048 (default):
https://pixeldrain.com/u/bScE8ANN
Dimension 2560:
https://pixeldrain.com/u/FoEbKSTL
-
Solista
- Posts: 202
- Joined: Thu Jul 06, 2023 11:47 am
- Location: Germany
- Has thanked: 7 times
- Been thanked: 293 times
- Contact:
Re: Stockfish dev sl
Marco Zerbitani drew my attention to a few things. I'll correct them as best I can. Here is my current answer to Marco:
Thank you for your factual arguments. I am also very interested in all differences being clarified as objectively as possible. That's why I think it's very good that you wrote to me privately.
Regarding the tutorial code: I'll fix that. I mentioned Eman style because, to my knowledge, Eman came first. If you like it I'll change my comments and write Eman & Sugar AI. You can also use and edit the learning file with Eman.
Unfortunately I made an error when uploading the files, that was not my intention. As a tool, there are actually two Sugar AI files in it without source code and comment. I will change that where I can. I will therefore also offer a complete SugarAI download.
Regarding Corchess: I may have made a mistake there too. Yes that's right, I didn't use the CorChess source code but used parts from Sugar AI (in terms of learning) and parts of the source code from the Stockfish developer. I. Ivec always implements the latest Stockfish patch into CorChess, while I implemented the CorChess source code into Stockfish developer. However, the executable EXE file should be identical. Maybe for the sake of accuracy I shouldn't refer to the engine name as CorChess and use a different name like it eg. A. Ces makes? He basically did the same thing but named his engine Cool Iris. Basically, Cool Iris is CorChess.
Here is Sugar AI. The learning code I use comes from this source code. The learning file can be edited (merged or defragmented) with the Sugar Experience Tool. The learning files from my engines can (renamed) also be used with Sugar AI and Eman (private engine by Khalid Omar).
Download: Sugar AI Special Edition
https://pixeldrain.com/u/BgjhRtAM
Thank you for your factual arguments. I am also very interested in all differences being clarified as objectively as possible. That's why I think it's very good that you wrote to me privately.
Regarding the tutorial code: I'll fix that. I mentioned Eman style because, to my knowledge, Eman came first. If you like it I'll change my comments and write Eman & Sugar AI. You can also use and edit the learning file with Eman.
Unfortunately I made an error when uploading the files, that was not my intention. As a tool, there are actually two Sugar AI files in it without source code and comment. I will change that where I can. I will therefore also offer a complete SugarAI download.
Regarding Corchess: I may have made a mistake there too. Yes that's right, I didn't use the CorChess source code but used parts from Sugar AI (in terms of learning) and parts of the source code from the Stockfish developer. I. Ivec always implements the latest Stockfish patch into CorChess, while I implemented the CorChess source code into Stockfish developer. However, the executable EXE file should be identical. Maybe for the sake of accuracy I shouldn't refer to the engine name as CorChess and use a different name like it eg. A. Ces makes? He basically did the same thing but named his engine Cool Iris. Basically, Cool Iris is CorChess.
Here is Sugar AI. The learning code I use comes from this source code. The learning file can be edited (merged or defragmented) with the Sugar Experience Tool. The learning files from my engines can (renamed) also be used with Sugar AI and Eman (private engine by Khalid Omar).
Download: Sugar AI Special Edition
https://pixeldrain.com/u/BgjhRtAM