It is currently Thu Jan 21, 2021 10:38 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 306 posts ]  Go to page Previous  1 ... 3, 4, 5, 6, 7, 8, 9 ... 16  Next
Author Message
Offline
 Post subject: Re: New android app "BadukAI"
Post #101 Posted: Mon Oct 05, 2020 8:42 am 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
BadukAI (KG 10b, 3s/move) - CSDL 5d 1:1 (CS W +R, KG W +R)
BadukAI (KG 10b, 4s/move) - CSDL 5d 1:1 (CS B +R, KG B +R)
BadukAI (KG 10b, 5s/move) - CSDL 5d 0:2 (CS B +R, CS W +R)


This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #102 Posted: Mon Oct 05, 2020 1:23 pm 
Dies with sente

Posts: 78
Liked others: 12
Was liked: 96
I made a new version (0.8), which fixes the bug discovered by @And (see post #97). Thanks again for the help in analyzing the error.


This post by akigo was liked by 3 people: Amigo, And, go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #103 Posted: Tue Oct 06, 2020 10:48 am 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
BadukAI (KG 10b, 8s/move) - CSDL 5d 3:1 (KG B +1.5, KG W +R, CS W +R, KG W +24.5)


This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #104 Posted: Tue Oct 06, 2020 1:04 pm 
Dies in gote

Posts: 34
Liked others: 18
Was liked: 7
akigo, thanks a lot for KataGo for android! will you add Chinese rules soon? they are somehow more familiar to people ... :) and you can make a window, like with a message about surrender, when KataGo passes? otherwise it’s not immediately clear whether this is a pass or something else.


This post by Amigo was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #105 Posted: Thu Oct 08, 2020 11:49 am 
Dies with sente

Posts: 78
Liked others: 12
Was liked: 96
I made a new version (0.9) with the following changes:

1) When AI passes, a popup message is displayed shortly, which informs about the pass
2) KataGo now uses chinese rules


This post by akigo was liked by 3 people: Amigo, And, go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #106 Posted: Thu Oct 08, 2020 11:50 am 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
BadukAI (KG 10b, 10s/move) - CSDL 5d 8:2 (time parity)


This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #107 Posted: Thu Oct 08, 2020 11:59 am 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
akigo wrote:
I made a new version (0.9) with the following changes:

1) When AI passes, a popup message is displayed shortly, which informs about the pass
2) KataGo now uses chinese rules

many many gratitude !!!


This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #108 Posted: Thu Oct 08, 2020 12:47 pm 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
lightvector, akigo
you make the world a better place and people happier! very pleasant program! :D


This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #109 Posted: Fri Oct 09, 2020 2:23 am 
Lives in gote

Posts: 322
Liked others: 16
Was liked: 94
And wrote:
BadukAI (KG 10b, 10s/move) - CSDL 5d 8:2 (time parity)
I don't understand... With my smartphone (Exynos 990) , BadukAI KG 20b, 10s/move usually wins (as W) at H6 against CSDL 5d, H6 seems a fair handicap. KG 20b, 10s/move is several ranks stronger. CSDL 5d can't win 20% of its games in even games.
I just tried 12 such even games, and predictably they all went rapidly the same way : after 50-70 moves, CSDL's winrate was always around 0. (For example, BadukAI is B in the 4 screenshots)
Attachment:
1.jpg
1.jpg [ 218.04 KiB | Viewed 708 times ]
How is it possible ? Is the 10b network so much weaker then the 20b ?


This post by Vargo was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #110 Posted: Fri Oct 09, 2020 5:17 am 
Dies with sente

Posts: 113
Liked others: 152
Was liked: 22
Rank: 25 kyu
I have the 6b net w/4 playouts (1 thread) stronger than gnugo (ogs 13k), the 10b policy network at around 3k, the 15b policy at 2-3 dan, the 20b policy at 4 dan, and the 40*256 bot policy at around 5 dan. The 10b at 9 payouts (1 thread) can beat the 15 bot policy. So it is a high kyu/low dan bot. The 20b is a high dan bot trained on 40b data. At 9 playouts, it is maybe 1 stone behind gtp4zen7 9dan. And what surprised me the most was that the 40b is the strongest at 5 sec/move on modern phones!

attached is komi 9 game with 10b policy (black) winning against pachi at default settings on my pentium 4405u. I cannot get pachi to play as black because of some bug, but the result of this game is typical between the two.


Attachments:
File comment: mislabeled, but nice game with 10b policy (black) winning against pachi at default settings on pentium 4405u komi 9
pachi.sgf [1.59 KiB]
Downloaded 10 times
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #111 Posted: Fri Oct 09, 2020 8:42 am 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
Vargo screenshots of two games, BadukAI (KG 10b, 10s/move) - CSDL 5d. sgf did not save


Attachments:
kg & cs.jpg
kg & cs.jpg [ 108.35 KiB | Viewed 662 times ]
kg & cs 2.jpg
kg & cs 2.jpg [ 108.02 KiB | Viewed 662 times ]

This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #112 Posted: Fri Oct 09, 2020 10:17 am 
Dies with sente

Posts: 78
Liked others: 12
Was liked: 96
At equal playouts the 20b network is about 1350 Elo stronger than the 10b network (see KataGo's README). At time parity we have to consider that the 10b network is about 8 times faster (at least on my device) due to the smaller size, so that the 10b network will have 8 times the number of playouts. Since doubling the playouts gains around 200 Elo (see https://github.com/leela-zero/leela-zero/issues/2236), this will reduce the gap by around 600 Elo. But we still have a whopping 750 Elo advantage for 20b at time parity. And the difference between @Vargo's and @And's configuration will be even larger, since @Vargo uses a faster device. So the difference in scores against CS is very plausible.


This post by akigo was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #113 Posted: Fri Oct 09, 2020 12:02 pm 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
BadukAI (KG 15b, 10 playouts) - CS Zero 6d 1:0 (BadukAI W +18.5)
BadukAI (KG 15b, 10 playouts) - CS Zero 7d 1:0 (BadukAI W +0.5)

BadukAI KG 15b, 10 playouts ~ 15 sec/move
Mobile Linpack 1.4 (https://play.google.com/store/apps/deta ... kbenchmark)
threads:2 max:9.33 mflop/sec, threads:1 max:7.78 mflop/sec
emulator nox

Vargo can you see the speed of your device mflop/sec?


Last edited by And on Fri Oct 09, 2020 1:07 pm, edited 1 time in total.

This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #114 Posted: Fri Oct 09, 2020 12:55 pm 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
CSDL 5d usually thinks for 10 seconds regardless of the device. launched on the emulator CS 5d (1 core) - CS 5d (4 cores) 3:7


This post by And was liked by: go4thewin
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #115 Posted: Fri Oct 09, 2020 10:25 pm 
Lives in gote

Posts: 322
Liked others: 16
Was liked: 94
@And

Hi, I use CSDL on my laptop (i7-9750H, 6 cores, 12 threads, Win 10)
At 5dan, one move takes around 0.1 or 0.2 sec. For example, below is a screenshot of a complete 205 moves game (CSDL5d v. CSDL5d).
Total time for the 205 moves was 29 sec (~0.14 sec/move)
Attachment:
scr.jpg
scr.jpg [ 113.5 KiB | Viewed 553 times ]
Curiously, on the Windows version of CSDL, the time shown is the total cumulated time, but in the Ipad version of CSDL, the time shown is the time for the current move only.
On my Ipad pro, at 5d, each move takes 5-10 sec, but the strength is the same as in the Win version.
Attachment:
ipad.jpg
ipad.jpg [ 429.73 KiB | Viewed 553 times ]

For the specifications of the Exynos 990 , see this site
I'll make some tests with KG10b and CSDL.

Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #116 Posted: Sat Oct 10, 2020 5:25 am 
Lives in gote

Posts: 322
Liked others: 16
Was liked: 94
Two games : BadukAI_KG_10b_10s/move v. CSDL 5d

At H6, BadukAI_KG10b loses, H6 seems too much.

At H4, BadukAI_KG10b wins


The winrate graph for the H4 game (Mmmmmmmh...the board is rotated 180° :scratch: )
Attachment:
h4.jpg
h4.jpg [ 78.27 KiB | Viewed 527 times ]


Attachments:
BadukAI10b10s_H4_CSDL5d.sgf [1.54 KiB]
Downloaded 82 times
BadukAI10b10s_H6_CSDL5d.sgf [2.23 KiB]
Downloaded 82 times
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #117 Posted: Sat Oct 10, 2020 6:32 am 
Dies with sente

Posts: 78
Liked others: 12
Was liked: 96
@And

just for comparison I ran the Mobile Linpack 1.4 on my device (Snapdragon 855+):

1 Thread: max. 999 mflops/s
2 Threads: max. 1030 mflops/s

And from geekbench 5 scores I know that @Vargo's Exynos 990 is still a little faster.


This post by akigo was liked by: And
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #118 Posted: Sat Oct 10, 2020 8:18 am 
Dies with sente

Posts: 113
Liked others: 152
Was liked: 22
Rank: 25 kyu
snapdragon 625 1 thread >150 (1st reading), <174 when device first turns on.
on this device 40*256 net is strongest at any time setting. Just to add another data point.

2nd device is Likebook mars e ink ereader processor=?. 1 thread <68.5 . Here, 20b takes maybe 4 seconds for 1 thread 1 playout policy move.


This post by go4thewin was liked by: And
Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #119 Posted: Sat Oct 10, 2020 8:52 am 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
akigo wrote:
@And

just for comparison I ran the Mobile Linpack 1.4 on my device (Snapdragon 855+):

1 Thread: max. 999 mflops/s
2 Threads: max. 1030 mflops/s

And from geekbench 5 scores I know that @Vargo's Exynos 990 is still a little faster.

Vargo

1 performance difference is more than 100!
2 engines work differently ("Crazy Stone has made a huge step forward by combining Deep Neural Networks with Monte Carlo Tree Search." https://play.google.com/store/apps/deta ... oid.gocsdl)
how does the strength of their game change with performance?
CSDL settings - 5d, and if its strength does not depend much on the device (it should not depend!) then everything is in order!
Vargo wrote:
...
On my Ipad pro, at 5d, each move takes 5-10 sec, but the strength is the same as in the Win version...

3 just a few games
given this, no results will surprise me! are you still surprised? :)

Top
 Profile  
 
Offline
 Post subject: Re: New android app "BadukAI"
Post #120 Posted: Sat Oct 10, 2020 12:51 pm 
Lives in gote
User avatar

Posts: 608
Liked others: 127
Was liked: 103
the program hangs at the end of the game. I could not get out of this state. saved sgf, does not react to anything else. if necessary, there is a log. this happened for the first time, does not bother, I inform you just in case


Attachments:
20201010-212744-C95B9D02.LOG.GZ [49.6 KiB]
Downloaded 9 times
freeze.jpg
freeze.jpg [ 156.21 KiB | Viewed 457 times ]


Last edited by And on Sat Oct 10, 2020 1:03 pm, edited 1 time in total.
Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 306 posts ]  Go to page Previous  1 ... 3, 4, 5, 6, 7, 8, 9 ... 16  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group