After a series of health issues over the last 3 years, I am back.
Since I am out of date:
o What is the preferred GUI/interface for engine testing?
o What is the preferred opening books for engine testing or is it a set of starting positions?
o What about eval tuning? I am aware of and understand all the NN theory and practice.
I am back but what is ....
Moderator: Ras
-
- Posts: 2091
- Joined: Mon Mar 13, 2006 2:31 am
- Location: North Carolina, USA
-
- Posts: 3220
- Joined: Wed Mar 10, 2010 10:18 pm
- Location: Hamburg, Germany
- Full name: Srdja Matovic
Re: I am back but what is ....
Welcome back Mr. Roberson:
https://www.chessprogramming.org/Cutechess-cli
nice one, built in SPRT:
https://www.talkchess.com/forum3/viewto ... =7&t=76370
https://github.com/official-stockfish/books
https://github.com/asdfjkl/neural_network_chess
https://github.com/asdfjkl/neural_netwo ... s/releases
Question might be:
- generate own datasets for training NNs or to use SF/Lc0 data sets?
https://github.com/glinscott/nnue-pytor ... g-datasets
- come up with own NN arch/trainer or use SF nnue-pytorch?
https://github.com/glinscott/nnue-pytorch
--
Srdja
Cutechess-cli
https://www.chessprogramming.org/Cutechess-cli
nice one, built in SPRT:
https://www.talkchess.com/forum3/viewto ... =7&t=76370
Idk for sure. IIRC SF started to use unbalanced UHO books?
https://github.com/official-stockfish/books
I like "Neural Networks for Chess" by Dominik Klein:
https://github.com/asdfjkl/neural_network_chess
https://github.com/asdfjkl/neural_netwo ... s/releases
Question might be:
- generate own datasets for training NNs or to use SF/Lc0 data sets?
https://github.com/glinscott/nnue-pytor ... g-datasets
- come up with own NN arch/trainer or use SF nnue-pytorch?
https://github.com/glinscott/nnue-pytorch
--
Srdja