Search found 527 matches

by corres
Mon Jan 28, 2019 3:30 pm
Forum: Computer Chess Club: General Topics
Topic: Distilled Networks for Lc0
Replies: 21
Views: 2042

Re: Distilled Networks for Lc0

OK, but these papers do not tell us your own works and the results depend on what you actually did. So I think an interpretation of your works would be needed. Naturally LC0 has the both head. But there are NN with separated value and policy head and there are NN in which the both are in the same s...
by corres
Mon Jan 28, 2019 8:57 am
Forum: Computer Chess Club: General Topics
Topic: LC0 bad search and poor tactics
Replies: 45
Views: 3436

Re: LC0 bad search and poor tactics

As I see there a lot of peoples who have been deluded the fast development of Leela and the obvious issues in the play of Leela. I should like to point to some facts: 1, The NN technique is about 40 years old, only the using in chess engine is a relative new thing. Mainly the last few years this tec...
by corres
Mon Jan 28, 2019 8:00 am
Forum: Computer Chess Club: General Topics
Topic: Distilled Networks for Lc0
Replies: 21
Views: 2042

Re: Distilled Networks for Lc0

There’s lots of opinion on knowledge distillation and academic papers to boot. Here’s a layperson appropriate article for starters: https://medium.com/neural-machines/knowledge-distillation-dc241d7c2322 OK, but these papers do not tell us your own works and the results depend on what you actually d...
by corres
Sun Jan 27, 2019 11:04 pm
Forum: Computer Chess Club: General Topics
Topic: Distilled Networks for Lc0
Replies: 21
Views: 2042

Re: Distilled Networks for Lc0

When you "distill" a network to get a smaller and faster NN it may lost some information. What is your experience about it? Did you make tests between original and "distilled" NN? How faster the "distilled" NN-s are? See the distilled network page for an extensive tournament testing various sizes. ...
by corres
Sun Jan 27, 2019 10:18 am
Forum: Computer Chess Club: General Topics
Topic: Distilled Networks for Lc0
Replies: 21
Views: 2042

Re: Distilled Networks for Lc0

What do you mean by “distilled”? You use one network to train another (usually smaller) network, rather than being trained by self-play or supervised learning data. When you "distill" a network to get a smaller and faster NN it may lost some information. What is your experience about it? Did you ma...
by corres
Sun Jan 27, 2019 9:16 am
Forum: Computer Chess Club: General Topics
Topic: Distilled Networks for Lc0
Replies: 21
Views: 2042

Re: Distilled Networks for Lc0

What do you mean by “distilled”? You use one network to train another (usually smaller) network, rather than being trained by self-play or supervised learning data. When you "distill" a network to get a smaller and faster NN it may lost some information. What is your experience about it? Did you ma...
by corres
Sat Jan 26, 2019 10:22 pm
Forum: Computer Chess Club: General Topics
Topic: LC0 bad search and poor tactics
Replies: 45
Views: 3436

Re: LC0 bad search and poor tactics

Yes, NN engines have a distinctly different style of play, however collectively software improvements have contributed more strength than one might think. Yes, hardware gets faster and certainly helps quite a lot, but software refinements in the aggregate are a major factor. For example: null move,...
by corres
Sat Jan 26, 2019 8:30 am
Forum: Computer Chess Club: General Topics
Topic: LC0 bad search and poor tactics
Replies: 45
Views: 3436

Re: LC0 bad search and poor tactics

Please don’t just say “lc0 did X.” Lc0 is an executable that can’t play chess on its own. It needs a network weights file to play. There are thousands of weights files, known by id numbers. If you refer to lc0, please also include the network id number. There are a few other families of weights fil...
by corres
Wed Jan 23, 2019 6:08 pm
Forum: Computer Chess Club: General Topics
Topic: Antifish vs Stockfish in cccc
Replies: 25
Views: 4168

Re: Antifish vs Stockfish in cccc

It seems this Leela Chess 1/2 really works well.
I would like to see a Stockfish Dev optimized to play against Leela...
by corres
Thu Jan 17, 2019 2:23 pm
Forum: Computer Chess Club: General Topics
Topic: Can somebody explain what makes Leela as strong as Komodo?
Replies: 50
Views: 4145

Re: Can somebody explain what makes Leela as strong as Komodo?

... But NN just seem to be an "natural" fit for gpus. Srdja Right and this is the essence. If NNs are such a natural fit for GPUs, why is NVIDIA's NN library so big? The Windows DLL is 280Mb, about 100x the size of the SF and LC0 binaries put together. It appears that NVIDIA have spent a huge amoun...