LcZero - CEDR Rating=3683 vBT4: Big transformer 4. New network architecture which builds off of BT3 by adding two types of auxiliary heads, future heads and categorical value heads. The categorical value heads predict a distribution over values of q rather than a WDL outcome distribution, and the future heads predict the moves that will be played over the next two plies. The hope is that these heads will give additional information to the net to improve training speed. We've also fixed half-precision training, so this model will be larger. BT4 training started in mid-October and is expected to take a few months. It has 15 layers with 1024 embedding size, 32 heads per layer, and 1536 emb size, for roughly a doubling in size over BT3. Individual statistics: Lc0 0.30.0 Stockfish 16 5/13 -3 13 Games Booot 7.2 6/10 +2 10 Games Critter 1.6a 8/9 +7 9 Games Dragon 3.2 4/8 +0 8 Games Stockfish 20230729 3.5/8 -1 8 Games ShashChess 33.2 3.5/8 -1 8 Games Fisherov chess monk 1.2 3/
Mars 1.0 os a strong engine. Lately I'm seeing there's a few new version releases such as v1.5, v1.6, v1.7 beta, and now this v1.8. Looks like they're all rated at 2993 elo. I wonder what the diffrences are among these versions if elo ratings are the same. Could it be playing style, etc.? Thanks.
ReplyDeleteRank refers to the latest version
DeleteHi Barbara,
DeleteI don't quite understand your reply. My question is about the differences among the versions since they're all rated 2993 elo.
Regards
Is there a seperate download site/git-hub, etc. ?
Deletepossible, but in my blog you download engine
DeleteIs there 32-bit for Mars 1.5 for Windows XP (non-AVX)?
ReplyDelete