Life In 19x19 http://lifein19x19.com/ |
|
KataGo v1.4 http://lifein19x19.com/viewtopic.php?f=18&t=17474 |
Page 2 of 4 |
Author: | ez4u [ Tue May 19, 2020 2:48 pm ] |
Post subject: | Re: KataGo v1.4 |
Bill Spight wrote: Dragon wrote: Imagine that 19x19 is solved. What do you think, KataGo (the best network, by default) makes the best move 1 time out of 5? out of 10? maybe a difficult question, but your opinion is interesting Well, if 19x19 go is solved, how was it solved? Surely by the best network... Isn't this a step too far? How would any net "solve" Go? I thought that the magic of nets is that they are able to play as strongly as they can despite (because?) being non-deterministic. So how will you ever be able to determine that you have solved Go if you are using a net? |
Author: | Dragon [ Tue May 19, 2020 3:33 pm ] |
Post subject: | Re: KataGo v1.4 |
I meant to imagine, assume that go 19x19 solved |
Author: | Bill Spight [ Tue May 19, 2020 3:43 pm ] |
Post subject: | Re: KataGo v1.4 |
ez4u wrote: Bill Spight wrote: Dragon wrote: Imagine that 19x19 is solved. What do you think, KataGo (the best network, by default) makes the best move 1 time out of 5? out of 10? maybe a difficult question, but your opinion is interesting Well, if 19x19 go is solved, how was it solved? Surely by the best network... Isn't this a step too far? How would any net "solve" Go? I thought that the magic of nets is that they are able to play as strongly as they can despite (because?) being non-deterministic. So how will you ever be able to determine that you have solved Go if you are using a net? I assumed that by "best network" Dragon meant the best program. I guess I should have said, "best program", myself. What shape future programs will take, who knows? I doubt if 19x19 go will ever be solved, anyway. However, if it is, and however it may be solved, the solution can be checked with a weaker program than that solved it. |
Author: | Dragon [ Tue May 19, 2020 4:00 pm ] |
Post subject: | Re: KataGo v1.4 |
I believe that go can be resolved. and not only 19x19. modern programs learn the "network". neural network (?). If someone comes up with a program that will learn and change themselves! then I’m pretty sure that after a while she will solve go! and provide a solution and evidence |
Author: | And [ Tue May 19, 2020 6:50 pm ] |
Post subject: | Re: KataGo v1.4 |
KataGo - first place! http://www.yss-aya.com/cgos/19x19/standings.html http://www.yss-aya.com/cgos/13x13/standings.html http://www.yss-aya.com/cgos/9x9/standings.html |
Author: | Uberdude [ Sun May 24, 2020 2:12 am ] |
Post subject: | Re: KataGo v1.4 |
Finally got round to installing the new version, after genconfig it's twice as fast as before (probably I didn't tune / adjust numthreads before), nice! |
Author: | Sneegurd [ Sun May 24, 2020 8:45 am ] |
Post subject: | Re: KataGo v1.4 |
Can someone please hint me - how can I play it 13x13 or 9x9 - I'm using Sabaki. Different command line and obviously different neural net files? Thx |
Author: | lightvector [ Sun May 24, 2020 8:47 am ] |
Post subject: | Re: KataGo v1.4 |
Have you tried just changing the board size in sabaki? KataGo uses one neural net for all sizes. |
Author: | Sneegurd [ Sun May 24, 2020 8:55 am ] |
Post subject: | Re: KataGo v1.4 |
lightvector wrote: Have you tried just changing the board size in sabaki? I have tried meanwhile, oh, I was not expecting it to be that simple - thank you
KataGo uses one neural net for all sizes. |
Author: | go4thewin [ Mon May 25, 2020 5:29 am ] |
Post subject: | Re: KataGo v1.4 |
Uberdude wrote: Finally got round to installing the new version, after genconfig it's twice as fast as before (probably I didn't tune / adjust numthreads before), nice! i dont know if its a coincidence, but the 40b speed has improved alot on the webapp version too! https://katago-gui.herokuapp.com/static ... obile.html For me, this means I can play a quick game anytime on my eink ereader/real board |
Author: | And [ Fri May 29, 2020 8:12 am ] |
Post subject: | Re: KataGo v1.4 |
here http://rusgo.org/2020/04/29/leela-zero-vs-katago/ talking about 40x384 network. Is she learning or has she failed? |
Author: | Uberdude [ Sat May 30, 2020 1:22 pm ] |
Post subject: | Re: KataGo v1.4 |
Kim Yoonyoung 8p (https://www.twitch.tv/kimyoonyoung) is playing a variable handicap series against KataGo running on beefy V100 hardware. She lost 2 games on 2 stones and 6.5 reverse komi and just now one on 3 stones, the "this group won't die" died after a very nice not working ladder KataGo played out to get some helper stones in sente. 2nd 3 stone starting now. Edit: she won the 2nd on 3 stones by 8.5 points. |
Author: | And [ Sun May 31, 2020 6:40 am ] |
Post subject: | Re: KataGo v1.4 |
can these games be downloaded in sgf? |
Author: | johnsmith [ Sun May 31, 2020 8:32 am ] |
Post subject: | Re: KataGo v1.4 |
And wrote: can these games be downloaded in sgf? Here are the games: https://online-go.com/player/787607/
|
Author: | Vargo [ Sat Jun 06, 2020 9:11 am ] |
Post subject: | Re: KataGo v1.4 |
KG 1.4.2 g170-b20c256x2-s4384473088-d968438914 v. KG 1.4.2 g170e-b20c256x2-s5055114240-d1149032340 (new experimental net from https://github.com/lightvector/KataGo/releases/tag/v1.4.2-nets) 30 games, 1s/move, gpu : 1x GTX1080, gogui-twogtp 1.5.1 The new net wins 19-11 (25 games by resignation, 5 games W+0.5 or B+0.5, no error, no duplicate game) PS. In another 30 game test between these two nets, the score was 16-14, but I had forgotten to use -alternate, so the new net was 30 times W... |
Author: | And [ Sat Jun 06, 2020 9:22 am ] |
Post subject: | Re: KataGo v1.4 |
would now be made 30 times black, there would be 60 games |
Author: | Vargo [ Sat Jun 06, 2020 9:28 pm ] |
Post subject: | Re: KataGo v1.4 |
KG 1.4.2 g170-b20c256x2-s4384473088-d968438914 v. KG 1.4.2 g170e-b20c256x2-s5055114240-d1149032340 (new experimental net from https://github.com/lightvector/KataGo/r ... 1.4.2-nets) After 60 more games (total of 90 games, each side 45 times B, 1s/move, 1x GTX1080) The new net wins 50-40 (55.6%) --------------------------------------- 30 games between KG 1.4.2 g170-b40c256x2-s4679779328-d1149909226 (new experimental) and KG 1.4.2 g170-b40c256x2-s3708042240-d967973220 The new net wins 16-14 , 22 games by resignation and 8 games(!) W+0.5 or B+0.5 1s/move, 1x GTX1080, GoGui-TwoGtp 1.5.1, no error, no duplicate game |
Author: | Vargo [ Sun Jun 07, 2020 2:13 am ] |
Post subject: | Re: KataGo v1.4 |
KG 1.4.2 g170-b30c320x2-s3530176512-d968463914 v. KG 1.4.2 g170-b30c320x2-s4432082944-d1149895217 (new experimental net) 50 games, 1s/move, 1x GTX1080 The experimental net wins 28-22 (42 games by resignation, 7 games B+0.5 or W+0.5 and 1 game by 1.5) |
Author: | Vargo [ Mon Jun 08, 2020 9:27 am ] |
Post subject: | Re: KataGo v1.4 |
new experimental nets 20b, 30b, 40b (from https://github.com/lightvector/KataGo/releases/tag/v1.4.2-nets) 50 game tests, TwoGtp 1.5.1, 1s/move, gpu : 1x GTX 1080 g170e-b20c256x2-s5055114240-d1149032340 v. g170-b30c320x2-s4432082944-d1149895217 20b wins 31-19 (62%) No error, no duplicate, 43 games by resignation, 7 games B+0.5 or W+0.5 g170e-b20c256x2-s5055114240-d1149032340 v. g170-b40c256x2-s4679779328-d1149909226 20b wins 29-21 (58%) No error, no duplicate, 41 games by resignation, 9 games B+0.5 or W+0.5 g170-b30c320x2-s4432082944-d1149895217 v. g170-b40c256x2-s4679779328-d1149909226 30b wins 26-24 (52%) No error, no duplicate, 45 games by resignation, 5 games B+0.5 or W+0.5 On this PC, with 1x GTX1080 1s/move corresponds to maxVisits =~ 700-800 for g170e-b20c256x2-s5055114240-d1149032340 |
Author: | go4thewin [ Tue Jun 09, 2020 11:39 am ] |
Post subject: | Re: KataGo v1.4 |
Is there anyway to play the nonzero version of katago on ogs? Thanks for the test matches! |
Page 2 of 4 | All times are UTC - 8 hours [ DST ] |
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |