TalkChess.com

1. Should continue as it is now

Firstly, people should be free to say what they please. Secondly, if having CTF keeps a bunch of the degenerates in their own space, all the better.

Good programs are easy to build

There really is nothing special about SF's NNUE. If you skip the half portion, it just a trivial NN, where the input layer is built with a sparse vector or matrix. I'de highly suggest anyone who is interested by it to roll their own implementation. Stockfish's trainer is ungodly to look at, and like...

2x Stockfish wins in a row? Next season Leela won't make it to the SuFi?

I'm sure there will be some changes, openings included, worry not Leela friends

Thanks for the test, and thank you for honoring my request to keep "SF-NNUE" in the name.

mwyoung wrote: ↑

Tue Oct 13, 2020 11:00 pm

And Stockfish as of today has shown itself lacking in good scaling.

That is an interesting take on a dominate TCEC SuFi victory, demonstrating that Leela connote scale on GPUs in the slightest.

Top 2? Hah, no. Its going to be 1. Stockfish 2. Komodo 3. Anyone in the top 10 who uses NNUE 3. Anyone in the top 10 who uses NNUE 3. Anyone in the top 10 who uses NNUE 4. Leela Yeah, right now Stockfish and Leela are the only projects that truely have the resources to train these Networks. But do ...

Top 2? Hah, no. Its going to be 1. Stockfish 2. Komodo 3. Anyone in the top 10 who uses NNUE 3. Anyone in the top 10 who uses NNUE 3. Anyone in the top 10 who uses NNUE 4. Leela Yeah, right now Stockfish and Leela are the only projects that truely have the resources to train these Networks. But do y...

ThatsIt wrote: ↑

Mon Oct 12, 2020 7:29 am

AndrewGrant wrote: ↑

Mon Oct 12, 2020 7:06 am

Thanks! Quite the lengthy name though

Maybe better that way: "Ethereal 12.75 x64 (PK+RPK-NN)" ?

Best wishes,
G.S.
(CEGT team)

Well you can just call it Ethereal 12.75

The PK and RPK NNs are compiled into the code. You can't turn them off lol.

Thanks! Quite the lengthy name though