Search found 9 matches

by dkappe
Wed Oct 31, 2018 5:20 pm
Forum: Computer Chess Club: General Topics
Topic: Leela Lite: A toolkit for experimenting with leela nets in python
Replies: 0
Views: 310

Leela Lite: A toolkit for experimenting with leela nets in python

For those wanting to experiment with Leela Chess Zero, but finding the highly optimized, multithreaded c++ code of lc0 too complex, I’ve released Leela Lite, a simple uct algorithm in python. It now has a script that allows it to be used as a uci engine. https://github.com/dkappe/leela_lite Comments...
by dkappe
Sun Oct 14, 2018 3:08 am
Forum: Computer Chess Club: General Topics
Topic: Official: Lc0 is the strongest engine :)
Replies: 40
Views: 6870

Re: Official: Lc0 is the strongest engine :)

It is actually one of the path to improvement. Dedicated Alpha / Beta "blunder checker", perhaps running on CPU which usually have spare cycles for LC0. I made one of these for leela in the days of the “blunder prone’ 192x15 networks. You can see the background and results: https://github.com/dkapp...
by dkappe
Thu Oct 11, 2018 4:36 pm
Forum: Computer Chess Club: General Topics
Topic: Official: Lc0 is the strongest engine :)
Replies: 40
Views: 6870

Re: Official: Lc0 is the strongest engine :)

You may want to use this tool to see how your gpu and cpu compare.

https://github.com/dkappe/leela-ratio
by dkappe
Wed Oct 10, 2018 3:12 am
Forum: Computer Chess Club: General Topics
Topic: Deep Pepper Chess
Replies: 1
Views: 623

Re: Deep Pepper Chess

It still doesn’t support uci, I presume? Also, any examples of play?
by dkappe
Wed Oct 10, 2018 2:58 am
Forum: Computer Chess Club: General Topics
Topic: Leela: Would it be reliable in correspondence chess?
Replies: 8
Views: 1815

Re: Leela: Would it be reliable in correspondence chess?

Sorry, the point being that it yields better, more useful results for analysis (and it may play a bit better than vanilla lc0).
by dkappe
Wed Oct 10, 2018 2:55 am
Forum: Computer Chess Club: General Topics
Topic: Leela: Would it be reliable in correspondence chess?
Replies: 8
Views: 1815

Re: Leela: Would it be reliable in correspondence chess?

I wasn’t too excited about lc0 multipv. It’s very easy for a single move to dominate the visits. I added a random root mod to my bender fork of lc0. After an initial 4000 iterations of search, it explores random root moves 1/16th of the time, then random root moves proportional to Q 3/16ths of the t...
by dkappe
Sat Sep 08, 2018 12:03 am
Forum: Computer Chess Club: General Topics
Topic: Lc0 v0.17.0-rc1 egt test
Replies: 12
Views: 2148

Re: Lc0 v0.17.0-rc1 egt test

With @dje's no history patch and no pruning, Ender 27 without TB trails only SF on a GTX 1070.

Code: Select all

Ender 27, Sergei.S.Markoff:Tablebase.Test - 5 sec, 2 threads, no prune
Success rate: 92.13% (363/394)
https://github.com/dkappe/leela-chess-w ... ndgame-Net
by dkappe
Tue Aug 28, 2018 3:34 am
Forum: Computer Chess Club: General Topics
Topic: Lc0 v0.17.0-rc1 egt test
Replies: 12
Views: 2148

Re: Lc0 v0.17.0-rc1 egt test

Thought I'd add two data points: on a GTX 1070, we have the following performance, without TB: Ender 6, Sergei.S.Markoff:Tablebase.Test - 5 sec, 2 threads Success rate: 85.53% (337/394) T947, Sergei.S.Markoff:Tablebase.Test - 5 sec, 2 threads Success rate: 75.38% (297/394) Note that the Leela nets d...