Thank you Lichess & Stockfish ♡
@revoof any speculation on when this will run on any iOS browser?
In the iOS Lichess app, I see “Stockfish 2021-02-20 Multi-Variant”, what SF version is running there?
thank you for stockfish 13 lichess!
I have a problem. everytime I use engine after closing tab and using it again it downloads the 20mb database again! it doesnt save it for the next visits
very good update. thank you very much lichess
Excellent work by the developers manually maintaining vectorised instructions. Excited to see SIMD SSE take over the open-source chess computing space for the masses.
I enabled the flag in Chrome 88 and downloaded the file, but I'm not exactly sure how to use this. I see the options on the right hand side such as: uci, go, stop, eval, etc. And I see a bunch of code generating on the screen, but I do not know how to use this interface.. Is there somewhere I can learn how to use this? Thanks in advance.
Hello? What means "NNUE"? I hate abbreviations...
Efficient Upgradable Neural Network.
I have questions in that regard:
1. Does StockFish "learn" from only analyzing our games ? or at least when we play against him here ?
2. Considering he does, is there a global knowledge repository / cloud for all Stockfish NNUE instances running over the world ? so as to have a global common knowledge. Not fragmented experience on several storages.
Stockfish NNUE doesn’t use Lichess games for training data. It generates them, I believe, by just having Stockfish play itself Chess until we have a good NNUE network: github.com/nodchip/Stockfish Training is computationally very expensive, which is why I still use Fairy-Stockfish-11.2 (pre-NNUE Stockfish with extensive Chess variant support) as my go to engine, especially for Chess variants.
It is Maia Chess that uses Lichess games for training data: maiachess.com/