I'm very happy to announce the release of Orion 1.0, a new version of my little chess engine !
This version includes:
a neural network trained "from zero", using only game results (1-0, 0-1, 1/2-1/2) and material as labels and targets ;
quantization of weights and biases, resulting in a 40% increase in nps (nodes per second) !
a completely rewritten Cerebrum library, allowing anyone to reproduce my work and obtain exactly the same neural network from the same data ;
a minor change, with a default Transposition Table size of 256 Mb (previously 128 Mb).
While the new version may be weaker than the previous one (approximately 100 elo), what mattered to me was no longer relying on evaluations from other engines for training data. I'm really happy to get such decent performance in a original and reproductible way !
Merci, David. Je m'étais mis à penser que vous aviez terminé ce projet. Eh, oui, espérons que la prochaine version d'Orion arrivera beaucoup plus vite que celle-ci!
David Carteau wrote: ↑Mon Mar 11, 2024 6:54 am
I'm very happy to announce the release of Orion 1.0, a new version of my little chess engine !
This version includes:
a neural network trained "from zero", using only game results (1-0, 0-1, 1/2-1/2) and material as labels and targets ;
quantization of weights and biases, resulting in a 40% increase in nps (nodes per second) !
a completely rewritten Cerebrum library, allowing anyone to reproduce my work and obtain exactly the same neural network from the same data ;
a minor change, with a default Transposition Table size of 256 Mb (previously 128 Mb).
While the new version may be weaker than the previous one (approximately 100 elo), what mattered to me was no longer relying on evaluations from other engines for training data. I'm really happy to get such decent performance in a original and reproductible way !