SFNNv9.5.3_240420 released...
SFNNv9.5.3_240420 released...
Update 240420 for Windows: Full updated to SF master and, more important, Mediumnet threshold changed to 1200. Old value of 300 is completely off! Pre-release
Both Windows versions of SFNNv9.5.3_240420 are released now with new default parameter for Mediumnet threshold of 1200. Old value of 300 was just a (dumb and wrong ) guess and completely off... For both Windows engines I got warnings from the stupid MS Defender. Just ignore it (like I do since the 10th false positive) or test it on Virus Total.
Use 1200 in SFNNv9.5.3_240410 or even make selfplay tests with different values to find a good MnTh parameter (to find the optimal one takes 10000s of games). Anyway, the optimal value of Mediumnet threshold (MnTh) depends (not that much on the same OS) on hardware and also on TC. Both dependencies could be minimized by empirical formulas, however 100000s of games are needed. Thus MnTh in the moment is an uci parameter.
However, with one(!) good MnTh value Stockfish dev can be easily beaten at STC and LTC. Here are the three tests with MnTh=1200 on my mini PC with Celeron N5095 (modern builds):
https://github.com/Joachim26/StockfishNPS/releases
Thank you Joachim26 !
Both Windows versions of SFNNv9.5.3_240420 are released now with new default parameter for Mediumnet threshold of 1200. Old value of 300 was just a (dumb and wrong ) guess and completely off... For both Windows engines I got warnings from the stupid MS Defender. Just ignore it (like I do since the 10th false positive) or test it on Virus Total.
Use 1200 in SFNNv9.5.3_240410 or even make selfplay tests with different values to find a good MnTh parameter (to find the optimal one takes 10000s of games). Anyway, the optimal value of Mediumnet threshold (MnTh) depends (not that much on the same OS) on hardware and also on TC. Both dependencies could be minimized by empirical formulas, however 100000s of games are needed. Thus MnTh in the moment is an uci parameter.
However, with one(!) good MnTh value Stockfish dev can be easily beaten at STC and LTC. Here are the three tests with MnTh=1200 on my mini PC with Celeron N5095 (modern builds):
https://github.com/Joachim26/StockfishNPS/releases
Thank you Joachim26 !
Re: SFNNv9.5.3_240420 released...
Thank you very much , Zamchess.
Re: SFNNv9.5.3_240420 released...
theres error in the engine it doesnt work
-
- Posts: 75
- Joined: Tue Nov 02, 2021 4:14 am
-
- Posts: 75
- Joined: Tue Nov 02, 2021 4:14 am
Re: SFNNv9.5.3_240420 released...
I borrowed the computer to make new compiles. It is not my decision the MS defender is there. I had to deactivate it for a while.Triton6654 wrote: ↑Sun Apr 21, 2024 1:14 pmInstead of MS Defender, U can use AVG Antivirus which is FREE n working well.
Re: SFNNv9.5.3_240420 released...
Many thanks for the update Zam!
-
- Posts: 632
- Joined: Thu Jun 10, 2010 12:50 am
Re: SFNNv9.5.3_240420 released...
Hello folks,
Does anyone know where the Menium network (20 MB, dimension 1024) comes from? Is this a newly trained network or is it an old SF network that Sf used in the first NNUE versions?
Ah I see from 23-12-02 23:41:58 What was this network trained for?
Does anyone know where the Menium network (20 MB, dimension 1024) comes from? Is this a newly trained network or is it an old SF network that Sf used in the first NNUE versions?
Ah I see from 23-12-02 23:41:58 What was this network trained for?
Peace be with you!
-
- Posts: 632
- Joined: Thu Jun 10, 2010 12:50 am
Re: SFNNv9.5.3_240420 released...
Does anyone know Joachim26's email? I can not find them.
https://github.com/Joachim26/StockfishNPS/releases
On lichess under BotLi I get an error message with the engine. Something is wrong in the UCI protocol.
Error message ... "OptimalTime". This does not exist in the UCI protocol.
I found this line in search.cpp:
//SFnps Begin
namespace SE = Stockfish::Eval;
int mediumNetTH = int(options["Mediumnet Threshold"]);
SE::tmOptTime = main_manager()->tm.optimum();
SE::mediumNetOn = (SE::tmOptTime < mediumNetTH);
std::cout << "OptimalTime " << SE::tmOptTime << sync_endl;
Maybe the line needs to read "OptimumTime" and not "OptimalTime"?
https://github.com/Joachim26/StockfishNPS/releases
On lichess under BotLi I get an error message with the engine. Something is wrong in the UCI protocol.
Error message ... "OptimalTime". This does not exist in the UCI protocol.
I found this line in search.cpp:
//SFnps Begin
namespace SE = Stockfish::Eval;
int mediumNetTH = int(options["Mediumnet Threshold"]);
SE::tmOptTime = main_manager()->tm.optimum();
SE::mediumNetOn = (SE::tmOptTime < mediumNetTH);
std::cout << "OptimalTime " << SE::tmOptTime << sync_endl;
Maybe the line needs to read "OptimumTime" and not "OptimalTime"?
Peace be with you!