News for hackers. Stuff that matters.
Re:Keras Implementation of Alpha Zero for Chess by bradknowles on Saturday March 03, 2018 @04:46PM (#16511802)
In the repo, they suggest that you instead go support another project. Namely, https://github.com/glinscott/leela-chess
Re:Keras Implementation of Alpha Zero for Chess by waterhouse on Sunday March 04, 2018 @03:13AM (#16513742)
Oh man, it's like BOINC. (In fact, perhaps one could incorporate it into BOINC.) I wonder if, after a while, the most powerful ML platforms in the world will be volunteer-run distributed systems. That would be an interesting situation.
Re:Keras Implementation of Alpha Zero for Chess by thope on Sunday March 04, 2018 @04:43AM (#16513977)
If anybody is interested, you will find ~300 millions PGN to train your NN models on chess games here : https://database.lichess.org/
Re:Keras Implementation of Alpha Zero for Chess by atroyn on Sunday March 04, 2018 @05:12AM (#16514063)
One interesting thing about the 'zero' approach is that it learns through self-play. So you can use a database to bootstrap the learning, but in principle you'd only need the game rules and enough training time.