Search found 25 matches

by jackd
Thu Feb 28, 2019 3:29 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

I just realized you can add a softmax directly after a convolutional layer. Before I was interpreting the policy head as (71 @ 3 * 3 * 256) -> batchnorm -> relu -> dense( 64 * 71 ) -> softmax But now I realize you apply the softmax directly to the 73 images. I am excited and expecting even better re...
by jackd
Fri Feb 22, 2019 3:01 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

You guys.. my policy network isn't that bad! After 100000 steps and a batch size of 2000 on a 10 x 128 it can beat me! I'm so excited I had to post something. I'll follow up with more detail eventually
by jackd
Sun Feb 17, 2019 5:00 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

It's worth mentioning that I wrote I was considering using momentum because I saw it was used by A0, not because I was getting good results with it. After realizing that batch normalization is what allowed A0 to use the parameters it did, I realized the correct choice for my network is RMSPROP until...
by jackd
Sun Feb 17, 2019 2:24 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

I did not filter but 80% of the games should have >2000 elo games. I am having difficulty getting a super strong net with only supervised learning so far though I still need to try few things like a) dropping learning rate after each epoch -- which already seems to help a bit. b) Ordering games by ...
by jackd
Sun Feb 17, 2019 1:07 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

@Daniel Shawul How strong is your policy network on it's own? It is weak tactically so can be exploited very easily. But its positional play is awesome, infact it can win games against tscp from time to time using only policy network (1 node mcts with 20x256 net). I don't know its strength but it a...
by jackd
Sat Feb 16, 2019 3:38 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

@Daniel Shawul How strong is your policy network on it's own? It is weak tactically so can be exploited very easily. But its positional play is awesome, infact it can win games against tscp from time to time using only policy network (1 node mcts with 20x256 net). I don't know its strength but it a...
by jackd
Fri Feb 15, 2019 9:34 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

@Daniel Shawul

How strong is your policy network on it's own?
by jackd
Fri Feb 15, 2019 8:06 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

Re: A0 subtlety

If the input is always oriented for the side to move, then should the policy output also be mirrored for black? Like if the best move was g8f6 the policy output would b1c3?
by jackd
Mon Feb 11, 2019 4:17 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: A0 subtlety
Replies: 29
Views: 2904

A0 subtlety

Was a set of input planes representing a position at time (t - T + 1) oriented for the side to move at time t or time (t - T + 1)?