| Life In 19x19 http://lifein19x19.com/ |
|
| KataGo Distributed Training and new networks http://lifein19x19.com/viewtopic.php?f=18&t=17944 |
Page 13 of 22 |
| Author: | And [ Tue May 24, 2022 7:12 am ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
LizGoban 0.7.0 https://github.com/kaorahi/lizgoban/releases |
|
| Author: | And [ Thu Jun 09, 2022 10:23 am ] | |||
| Post subject: | Re: KataGo Distributed Training and new networks | |||
In the latest version of AhQ Go lite (3.3.0), a 4 block network is used to play at the 1-3 dan level! I think this is an interesting alternative to Calibrated Rank Bot katrain! katago 4b 10 playouts - gtp4zen 1d:
|
||||
| Author: | And [ Thu Jun 09, 2022 10:30 am ] | ||
| Post subject: | Re: KataGo Distributed Training and new networks | ||
the same, katago 4b 100 playouts - gtp4zen 3d:
|
|||
| Author: | And [ Thu Jun 09, 2022 10:34 am ] | ||
| Post subject: | Re: KataGo Distributed Training and new networks | ||
the same, katago 4b 500 playouts - gtp4zen 5d:
|
|||
| Author: | And [ Fri Jun 10, 2022 9:06 am ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
unfortunately, this 4 block network often makes serious mistakes. p.s. it seems to me that kyu-rank bot BadukAI has become much better at playing! |
|
| Author: | And [ Sun Jun 12, 2022 9:52 am ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
KaTrain 7d - CS Zero 7d 0:2 (B+, W+)  kata1-b60c320-s594 |
|
| Author: | And [ Wed Jun 15, 2022 2:58 am ] | ||
| Post subject: | Re: KataGo Distributed Training and new networks | ||
KaTrain 7d - CS Zero 7d 1:1 kata1-b60c320-s606
|
|||
| Author: | And [ Wed Jun 15, 2022 12:20 pm ] | ||
| Post subject: | Re: KataGo Distributed Training and new networks | ||
katago 4b 1 playouts - CS Zero 1k (g170small-b4c32-s709)
|
|||
| Author: | And [ Sat Jun 18, 2022 11:29 pm ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
Sayuri https://github.com/CGLemon/Sayuri "Let's ROCK! Sayuri is a GTP-compliant go engine which supports variable komi and board size. Strongly inspired by Leela Zero and Kata Go. Based on Deep Convolutional Neural Network, Monte Carlo Tree Search and other powerful technics." |
|
| Author: | Amigo [ Sat Jun 25, 2022 1:49 am ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
And wrote: Sayuri https://github.com/CGLemon/Sayuri "Let's ROCK! Sayuri is a GTP-compliant go engine which supports variable komi and board size. Strongly inspired by Leela Zero and Kata Go. Based on Deep Convolutional Neural Network, Monte Carlo Tree Search and other powerful technics." does it work? I can't even find the network. |
|
| Author: | And [ Mon Jun 27, 2022 12:09 am ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
what are the results of 20b - 40b time parity? didn't find any games 
|
|
| Author: | And [ Mon Jun 27, 2022 12:24 pm ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
Sayuri - network can be downloaded |
|
| Author: | And [ Mon Jun 27, 2022 9:57 pm ] | |||
| Post subject: | Re: KataGo Distributed Training and new networks | |||
katago b40c256-s118 50 playouts (B) - b20c256x2-s530 100 playouts (W) 1 and 2 game 0:2 (TIME_B: 2536.3 TIME_W: 2364) (TIME_B: 1985 TIME_W: 1656.5)
|
||||
| Author: | And [ Thu Jun 30, 2022 12:25 am ] | |||
| Post subject: | Re: KataGo Distributed Training and new networks | |||
3 and 4 game 0:2 (0:4) (TIME_B: 2301.1 TIME_W: 1822.5) (TIME_B: 2373.6 TIME_W: 1905.4)
|
||||
| Author: | And [ Fri Jul 01, 2022 1:47 am ] | |||
| Post subject: | Re: KataGo Distributed Training and new networks | |||
5 and 6 game 1:1 (1:5) (TIME_B: 2090 TIME_W: 1525.9) (TIME_B: 1612.1 TIME_W: 1903.2)
|
||||
| Author: | And [ Mon Jul 04, 2022 7:14 am ] | |||
| Post subject: | Re: KataGo Distributed Training and new networks | |||
7 and 8 game 2:0 (3:5) (TIME_B: 1825.3 TIME_W: 1867.4) (TIME_B: 1902.4 TIME_W: 1969.4)
|
||||
| Author: | lightvector [ Mon Jul 04, 2022 1:57 pm ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
When testing such small number of playouts, take care that numSearchThreads is set to a very small number too. I would recommend to set the number of search threads to no more than 1/20th of the number of playouts. If you have too many threads competing to perform very few playouts while the search tree is very small, many threads will be forced to search garbage variations, which could result in even worse play than having fewer threads and using fewer playouts. |
|
| Author: | And [ Tue Jul 05, 2022 1:22 am ] |
| Post subject: | Re: KataGo Distributed Training and new networks |
this is useful information, thank you very much! ps in these games, the number of search threads = 3 |
|
| Author: | And [ Wed Jul 06, 2022 5:41 am ] | |||
| Post subject: | Re: KataGo Distributed Training and new networks | |||
9 and 10 game 1:1 (TIME_B: 1929.8 TIME_W: 2416.8) (TIME_B: 2099.1 TIME_W: 1584.4) after 10 games katago b40c256-s118 50 playouts (B) - b20c256x2-s530 100 playouts (W) 4:6
|
||||
| Author: | And [ Wed Jul 06, 2022 5:53 am ] | ||
| Post subject: | Re: KataGo Distributed Training and new networks | ||
katago b40c256-s118 50 playouts (W) - b20c256x2-s530 100 playouts (B) 9:1 after 20 games katago b40c256-s118 50 playouts - b20c256x2-s530 100 playouts 13:7 b20 (B) won the 7th game (TIME_B: 1791 TIME_W: 2049.1)
|
|||
| Page 13 of 22 | All times are UTC - 8 hours [ DST ] |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|
