It is currently Tue Jan 22, 2019 5:06 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 266 posts ]  Go to page Previous  1 ... 8, 9, 10, 11, 12, 13, 14  Next
Author Message
Offline
 Post subject: Re: LZ's progression
Post #201 Posted: Sat Nov 03, 2018 1:18 pm 
Dies in gote

Posts: 69
Liked others: 2
Was liked: 9
Rank: KGS 2 D
There is a well-known saying among professional go players. When you have to kill a huge dragon to win, you are already far behind. It may still take more time for LZ to understand this. When LZ186 plays black, it almost always tries to kill a huge dragon. LZ185 is more modest and plays more defensive moves.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #202 Posted: Sat Nov 03, 2018 1:42 pm 
Dies in gote

Posts: 29
Liked others: 0
Was liked: 2
splee99 wrote:
... When LZ186 plays black, it almost always tries to kill a huge dragon. LZ185 is more modest and plays more defensive moves.

And what about #157a? It have the same results, so i'm desagree with you.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #203 Posted: Sat Nov 03, 2018 3:27 pm 
Dies in gote

Posts: 69
Liked others: 2
Was liked: 9
Rank: KGS 2 D
LZ157 is using 15 blocks, so it is unable to generate any huge dragon (at least as large as the one generated by the 40 block). So there is less unexpected outcome regarding life and death of a large dragon.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #204 Posted: Sat Nov 03, 2018 3:40 pm 
Judan

Posts: 5462
Location: Cambridge, UK
Liked others: 306
Was liked: 2890
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Yakago wrote:
luigi wrote:
I'm following Leela's progression with delight. Does anyone know what its current rating against human pros on "normal" hardware (let's say one GTX 1080) would be? Is it better than the #1 human on those conditions? If so, by how much?


Considering that even a few months ago, LeelaZero was beating human professionals with 2-3 handicap stones, which it was not even built for, on consumer hardware. (Since then, there has been various improvements to handicap play)

I would think that LeelaZero is well above any human on a GTX 1080. It is possible that you can find a blind spot with a ladder or similar though, depending on the time settings.


I agree LZ is now super-human on a GTX 1080 with say 30s a move, but I'm not actually aware of much direct evidence. Yakago, do you refer to Haylee, she is a lot weaker than a top pro. But yes there's lots of transitive evidence, e.g. a while ago LZ bijxo got some wins in the match vs Golaxy and Golaxy spanked all the humans it played (except that one loss). Reviewing pro games with LZ (and comparing with Elf) it appears to me that LZ is stronger and much more consistent, though it is of course possible the pro had a good response LZ didn't see that invalidates LZ's judgement, but I expect that's a minority (and I can find some, particularly ladders). Or on wbaduk I see LZ beating up non-top Japanese pros, often winning by resign in under 100 moves.

I did make an account on Fox to play with LZ to try to play some of the 9ds there and hopefully eventually top pros to gather some direct evidence, but as you start at 3d and need to win 20 games in a row to double rank promote that's 60 games of grinding to 9d, plus I felt a bit bad about beating up 3ds even though my username says LeelaZero and I say I'm a bot so they can quit before the game counts if they don't want to play (but most are Chinese so language barrier). I did actually play a few games this morning, this one below I liked move 47 armpit hit kosumi approach to 4-4 (that's usually a terrible beginner mistake) as a leaning attack to kill the group above, though of course my opp helped with that unnecessary c2 defence.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #205 Posted: Sun Nov 04, 2018 1:25 am 
Lives with ko
User avatar

Posts: 250
Liked others: 24
Was liked: 72
Some games of LZ#131 (network ecab83bb, 192x15) against Haylee last May were on relatively powerful hardware. The comments in https://online-go.com/game/12760703 indicate that at each move, the number of visits was typically 200000. On the other hand, https://online-go.com/game/12665694 indicates "1x1080Ti", and a number of visits less than 100000 in general, which is more reasonable.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #206 Posted: Sun Nov 04, 2018 7:02 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
20 game match between LZ0.16 #187 and Elf v1
(1x1080, 5 min per side and game, no pondering, twogtp 1.4.10)

Elf v1 wins 13:7 (7 wins as B, 6 wins as W)

#187 won 35% of its games, not so bad ;-)

(reminder : #184 had won only 30% of its 20 games against Elf v1)
Maybe not significant, but it seems to go in the right direction)

Average length : 220 moves , average time used per side and game: 215" (about the same for both sides)
Attachment:
elf2_187_ELFisW.zip [8.27 KiB]
Downloaded 20 times
Attachment:
elf2_187-ELFisB.zip [9.1 KiB]
Downloaded 14 times

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #207 Posted: Sun Nov 04, 2018 1:56 pm 
Lives with ko

Posts: 219
Liked others: 160
Was liked: 117
splee99 wrote:
There is a well-known saying among professional go players. When you have to kill a huge dragon to win, you are already far behind. It may still take more time for LZ to understand this. When LZ186 plays black, it almost always tries to kill a huge dragon. LZ185 is more modest and plays more defensive moves.


Recent Go AIs are known for not paying much attention to human pros' sayings, but rather for proving humans wrong :-)

_________________
Sorin - 361points.com

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #208 Posted: Sun Nov 04, 2018 4:55 pm 
Dies in gote

Posts: 69
Liked others: 2
Was liked: 9
Rank: KGS 2 D
Well, overplay is always wrong. As we can see LZ186 is well punished by LZ187 which is quite often in the defensive mode.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #209 Posted: Sun Nov 04, 2018 10:39 pm 
Dies in gote

Posts: 29
Liked others: 0
Was liked: 2
Vargo wrote:
20 game match between LZ0.16 #187 and Elf v1
#187 won 35% of its games, not so bad ;-)

It looks like you are an optimist :lol: :
Code:
# Black: Leela Zero
# BlackCommand: C:\APPS\l0gpu16\leelaz.exe --gtp --weights=C:\APPS\net\6c7c1c83.gz -t 12 --gpu 0 --gpu 1 --noponder --precision single
# BlackLabel: Leela Zero:0.16
# BlackVersion: 0.16
# Date: November 5, 2018 5:10:21 AM VLAT
# Host: DESKTOP-SP9U1O5
# Komi: 7.5
# Referee: -
# Size: 19
# White: Leela Zero
# WhiteCommand: C:\APPS\l0gpu16\leelaz.exe --gtp --weights=C:\APPS\net\d13c4099.gz -t 12 --gpu 0 --gpu 1 --noponder --precision single
# WhiteLabel: Leela Zero:0.16
# WhiteVersion: 0.16
# Xml: 0
#
#GAME   RES_B   RES_W   RES_R   ALT   DUP   LEN   TIME_B   TIME_W   CPU_B   CPU_W   ERR   ERR_MSG
0   W+R   W+R   W+R   0   -   228   352   344.5   0   0   0   
1   W+R   W+R   W+R   0   -   244   378.8   369.4   0   0   0   
2   B+R   B+R   B+R   0   -   221   340.8   336   0   0   0   
3   W+R   W+R   W+R   0   -   188   293   284.4   0   0   0   
4   B+R   B+R   B+R   0   -   261   402   396.4   0   0   0   
5   W+R   W+R   W+R   0   -   240   373   363.4   0   0   0   
6   W+R   W+R   W+R   0   -   238   370.5   360.4   0   0   0   
7   B+R   B+R   B+R   0   -   219   338.5   332.7   0   0   0   
8   W+R   W+R   W+R   0   -   258   400.4   390.6   0   0   0   
9   W+R   W+R   W+R   0   -   292   452.3   442.7   0   0   0   
10   W+R   W+R   W+R   0   -   278   430   420.3   0   0   0   
11   W+R   W+R   W+R   0   -   292   451.9   442.4   0   0   0   
12   W+R   W+R   W+R   0   -   226   347.8   342   0   0   0   
13   W+R   W+R   W+R   0   -   298   459.6   451   0   0   0   
14   W+R   W+R   W+R   0   -   250   386.4   379.2   0   0   0   
15   W+R   W+R   W+R   0   -   226   350.1   341.4   0   0   0   
16   W+R   W+R   W+R   0   -   248   385.8   375.3   0   0   0   
17   B+R   B+R   B+R   0   -   113   175   172   0   0   0   
18   W+R   W+R   W+R   0   -   314   485.5   475.3   0   0   0   
19   B+R   B+R   B+R   0   -   103   159.1   156.4   0   0   0   
20   W+R   W+R   W+R   0   -   210   326.3   317.6   0   0   0   
21   W+R   W+R   W+R   0   -   266   413.1   403.2   0   0   0   
22   W+R   W+R   W+R   0   -   240   371.1   363.3   0   0   0   
23   W+R   W+R   W+R   0   -   202   312.2   305.3   0   0   0   
24   W+R   W+R   W+R   0   -   246   382.1   371.9   0   0   0   
25   B+R   B+R   B+R   0   -   121   186.8   184   0   0   0   
26   W+R   W+R   W+R   0   -   264   407   398.5   0   0   0   
27   W+R   W+R   W+R   0   -   312   481.4   472.5   0   0   0   
28   W+R   W+R   W+R   0   -   210   326.6   317.7   0   0   0   
29   W+R   W+R   W+R   0   -   292   450.8   442.4   0   0   0   
30   B+R   B+R   B+R   0   -   91   141.2   138.7   0   0   0   
31   B+R   B+R   B+R   0   -   91   141.1   138.8   0   0   0   
32   W+R   W+R   W+R   0   -   288   448.3   435.7   0   0   0   
33   W+R   W+R   W+R   0   -   250   388.4   377.2   0   0   0   
34   W+R   W+R   W+R   0   -   220   341.3   332.3   0   0   0   
35   W+R   W+R   W+R   0   -   190   293.6   287.6   0   0   0   
36   B+R   B+R   B+R   0   -   189   290.7   287.1   0   0   0   
37   B+R   B+R   B+R   0   -   227   350.1   344.3   0   0   0   
38   W+R   W+R   W+R   0   -   362   558.6   548.2   0   0   0   
39   W+R   W+R   W+R   0   -   184   285.7   277.6   0   0   0   
40   W+R   W+R   W+R   0   -   174   270.8   263   0   0   0   
41   W+R   W+R   W+R   0   -   272   421.2   413   0   0   0   
42   W+R   W+R   W+R   0   -   298   458.1   452.1   0   0   0   
43   W+R   W+R   W+R   0   -   192   298.3   290.1   0   0   0   
44   W+R   W+R   W+R   0   -   196   304   296.1   0   0   0   
45   W+R   W+R   W+R   0   -   364   560.9   551.9   0   0   0   
46   W+R   W+R   W+R   0   -   280   431.2   424.2   0   0   0   
47   W+R   W+R   W+R   0   -   340   524.3   514.9   0   0   0   
48   B+R   B+R   B+R   0   -   271   417   411.4   0   0   0   
49   W+R   W+R   W+R   0   -   128   199.9   192.7   0   0   0

#187(black) vs Elf v1. : +11-39

P.S. 3s per move, 2x1080ti
P.S.S Games ##30-31 are elf's ladder-blindness, so everything is very bad. :shock:

EDIT.

#187(white) vs Elf v1. : +17-33
Code:
# Black: Leela Zero
# BlackCommand: C:\APPS\l0gpu16\leelaz.exe --gtp --weights=C:\APPS\net\d13c4099.gz -t 12 --gpu 0 --gpu 1 --noponder --precision single
# BlackLabel: Leela Zero:0.16
# BlackVersion: 0.16
# Date: November 5, 2018 3:14:25 PM VLAT
# Host: DESKTOP-SP9U1O5
# Komi: 7.5
# Referee: -
# Size: 19
# White: Leela Zero
# WhiteCommand: C:\APPS\l0gpu16\leelaz.exe --gtp --weights=C:\APPS\net\6c7c1c83.gz -t 12 --gpu 0 --gpu 1 --noponder --precision single
# WhiteLabel: Leela Zero:0.16
# WhiteVersion: 0.16
# Xml: 0
#
#GAME   RES_B   RES_W   RES_R   ALT   DUP   LEN   TIME_B   TIME_W   CPU_B   CPU_W   ERR   ERR_MSG
0   B+R   B+R   B+R   0   -   277   421.7   425.3   0   0   0   
1   B+R   B+R   B+R   0   -   231   352.1   354.7   0   0   0   
2   W+R   W+R   W+R   0   -   272   417   415.7   0   0   0   
3   B+R   B+R   B+R   0   -   217   331   334.1   0   0   0   
4   W+R   W+R   W+R   0   -   312   476.2   480   0   0   0   
5   W+R   W+R   W+R   0   -   128   197.3   195.7   0   0   0   
6   W+R   W+R   W+R   0   -   224   343.6   344.8   0   0   0   
7   B+R   B+R   B+R   0   -   285   434.5   439.2   0   0   0   
8   B+R   B+R   B+R   0   -   167   255.1   256.5   0   0   0   
9   B+R   B+R   B+R   0   -   229   349.6   351.5   0   0   0   
10   B+R   B+R   B+R   0   -   193   294.9   296   0   0   0   
11   B+R   B+R   B+R   0   -   213   324.6   327   0   0   0   
12   B+R   B+R   B+R   0   -   213   325.3   326.7   0   0   0   
13   B+R   B+R   B+R   0   -   303   462.2   464.1   0   0   0   
14   W+R   W+R   W+R   0   -   232   355.1   353.2   0   0   0   
15   B+R   B+R   B+R   0   -   181   275.8   277.6   0   0   0   
16   W+R   W+R   W+R   0   -   120   184.9   183.5   0   0   0   
17   W+R   W+R   W+R   0   -   212   324.7   324.3   0   0   0   
18   B+R   B+R   B+R   0   -   167   255.2   257.3   0   0   0   
19   W+R   W+R   W+R   0   -   172   264.4   263.9   0   0   0   
20   B+R   B+R   B+R   0   -   337   513.1   516.3   0   0   0   
21   B+R   B+R   B+R   0   -   319   486.2   488.6   0   0   0   
22   B+R   B+R   B+R   0   -   175   267.4   269.2   0   0   0   
23   B+R   B+R   B+R   0   -   209   319   322.4   0   0   0   
24   B+R   B+R   B+R   0   -   289   440.3   445.6   0   0   0   
25   W+R   W+R   W+R   0   -   142   218.8   216.2   0   0   0   
26   B+R   B+R   B+R   0   -   313   475.8   480.9   0   0   0   
27   B+R   B+R   B+R   0   -   337   513.3   517.5   0   0   0   
28   B+R   B+R   B+R   0   -   305   465.2   466   0   0   0   
29   W+R   W+R   W+R   0   -   146   225.3   222.9   0   0   0   
30   W+R   W+R   W+R   0   -   164   251.9   249.3   0   0   0   
31   W+R   W+R   W+R   0   -   158   242.7   240.5   0   0   0   
32   W+R   W+R   W+R   0   -   122   187.8   185.7   0   0   0   
33   B+R   B+R   B+R   0   -   245   373.5   375.7   0   0   0   
34   B+R   B+R   B+R   0   -   285   433.9   436.6   0   0   0   
35   B+R   B+R   B+R   0   -   161   245.9   247.4   0   0   0   
36   W+R   W+R   W+R   0   -   244   373   373   0   0   0   
37   W+R   W+R   W+R   0   -   192   295.1   293.9   0   0   0   
38   B+R   B+R   B+R   0   -   291   443.2   447.4   0   0   0   
39   B+R   B+R   B+R   0   -   153   233.1   234.7   0   0   0   
40   W+R   W+R   W+R   0   -   172   264   263.9   0   0   0   
41   B+R   B+R   B+R   0   -   163   249.3   250.4   0   0   0   
42   B+R   B+R   B+R   0   -   247   376.8   378.8   0   0   0   
43   B+R   B+R   B+R   0   -   221   336.3   339.7   0   0   0   
44   W+R   W+R   W+R   0   -   100   155.3   151   0   0   0   
45   B+R   B+R   B+R   0   -   301   457.4   462.6   0   0   0   
46   B+R   B+R   B+R   0   -   345   524.7   533.3   0   0   0   
47   B+R   B+R   B+R   0   -   359   546.1   550.6   0   0   0   
48   B+R   B+R   B+R   0   -   299   454.5   458.5   0   0   0   
49   B+R   B+R   B+R   0   -   237   360.9   363.9   0   0   0

In total:

#187 vs Elf v1. : +28-72

P.S. If someone wants the games...

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #210 Posted: Mon Nov 05, 2018 7:17 pm 
Dies in gote

Posts: 29
Liked others: 0
Was liked: 2
On the other hand ("visit" parity test (-v 1601 -r 5)):
Code:
The first net is worse than the second
Elfv1 v #187 ( 74 games)
        wins        black       white
Elfv1   24 32.43%   15 34.88%    9 29.03%
#187    50 67.57%   28 65.12%   22 70.97%
                    43 58.11%   31 41.89%

C:\APPS\l0gpu16\validation.exe -n C:\APPS\net\d13c4099.gz -o "-g -v 1601 --gpu 0 --gpu 1 --noponder -t 12 -q -d --timemanage off --precision single -r 5 -w" -n C:\APPS\net\6c7c1c83.gz -o "-g -v 1601 --gpu 0 --gpu 1 --noponder -t 12 -q -d --timemanage off --precision single -r 5 -w" -- C:\APPS\l0gpu16\leelaz -- C:\APPS\l0gpu16\leelaz -k elfv1-187


Attachments:
elfv1-187-visit.zip [60.1 KiB]
Downloaded 17 times
Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #211 Posted: Fri Nov 30, 2018 6:19 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
Ten 10 game matches at time parity :

157 v 158
157 v 160
157 v 165
157 v 170
157 v 175
157 v 180
157 v 185
157 v 190
157 v 193

All games and results :
Attachment:
157vxxx_2m.zip [83.8 KiB]
Downloaded 15 times

Below, a graph of the win percentages of #157, with a dashed linear fit, showing the "average progression" of the networks.
For example, second black dot from left means : #157 has a win rate of 90% against #160

(1x1080, 2 min per side per game, no pondering, komi 7.5, twogtp 1.5.0, LZ016, #157 is always W in the even numbered games)
The networks get better indeed :tmbup: !
Attachment:
157vxxx.gif
157vxxx.gif [ 11.1 KiB | Viewed 1412 times ]


I'm running the exact same experiment a second time...Results this evening or tomorrow.

I hope the linear fits will look similar.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #212 Posted: Fri Nov 30, 2018 11:25 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
The second series of matches shows progress too, but not as much as the first one.

(Same commands, same hardware)
Games and results of the second series:
Attachment:
157_xxxxx.zip [82.84 KiB]
Downloaded 13 times


The combined graph (ten 20-game matches):
#157 average win rate goes down from ~75% to ~45%
Attachment:
157vxxx_2.gif
157vxxx_2.gif [ 10.79 KiB | Viewed 1356 times ]

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #213 Posted: Sat Dec 01, 2018 3:05 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
Third and last experiment with another ten 10-game matches (I should have done the three experiments in one go, sorry...)
Attachment:
157vxxxxx.zip [79.43 KiB]
Downloaded 12 times

After these 270 matches at short time parity (~1 sec/move with 1x1080), the combined graph shows real progress for the 40x256 networks. I think they are now better than #157 (last 15x192), even for relatively fast games.

I've not plotted the linear fit, because I'm sure someone will tell me the long time progression model is not linear (true, but it can be a good approximation, locally)
For example : rightmost blue point means #157 wins 40% against #193
Attachment:
157vxxx_3.gif
157vxxx_3.gif [ 8.12 KiB | Viewed 1265 times ]

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #214 Posted: Sun Dec 09, 2018 5:13 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
20 game match :#194 v. #157 at time parity
5min per side per game, komi 7.5, -r 10 for both, LZ 0.16, twogtp 1.5.0
PONDERING ENABLED for both (#194 using gpu 0, and #157 using gpu 1 , gpu : 2x1080Ti)

#194 wins 12-8 (7 wins as W)

If someone wants the games or the stats, I'll upload them

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #215 Posted: Sun Dec 09, 2018 1:58 pm 
Dies in gote

Posts: 69
Liked others: 2
Was liked: 9
Rank: KGS 2 D
Another strong opponent is the Leela master. I tested the 15b LM_GX88 which is indeed stronger than #193 in time parity games. However #194 is now much stronger than LM_GX88.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #216 Posted: Mon Dec 10, 2018 3:19 pm 
Dies in gote

Posts: 29
Liked others: 0
Was liked: 2
Forgotten test of "home-made" LeelaMaster_GX88 vs #157a net.

By visits (standart 3200):
Code:
The first net is better than the second
LeelaMaster_GX88.txt v fc5e0a50.gz ( 231 games)
                          wins        black       white
LeelaMaster_GX88.txt  136 58.87%   58 61.05%   78 57.35%
fc5e0a50.gz            95 41.13%   37 38.95%   58 42.65%
                                   95 41.13%  136 58.87%

C:\APPS\l0gpu16\validation.exe -n C:\APPS\net\LeelaMaster_GX88.txt -o "-g -v 3200 --gpu 0 --gpu 1 --noponder -t 12 -q -d --timemanage off --precision single -w" -n C:\APPS\net\fc5e0a50.gz -o "-g -v 3200 --gpu 0 --gpu 1 --noponder -t 12 -q -d --timemanage off --precision single -w" -- C:\APPS\l0gpu16\leelaz -- C:\APPS\l0gpu16\leelaz -k lmgx88-157a


By time-parity (3s per move):
Code:
The first net is worse than the second
LeelaMaster_GX88.txt v fc5e0a50.gz ( 308 games)
                          wins        black       white
LeelaMaster_GX88.txt  147 47.73%   53 46.90%   94 48.21%
fc5e0a50.gz           161 52.27%   60 53.10%  101 51.79%
                                  113 36.69%  195 63.31%

C:\APPS\l0gpu16\validation.exe -n C:\APPS\net\LeelaMaster_GX88.txt -o "-g --gpu 0 --gpu 1 --noponder -t 12 -d --precision single --logfile log -w" -n C:\APPS\net\fc5e0a50.gz -o "-g --gpu 0 --gpu 1 --noponder -t 12 -d --precision single --logfile log -w" -- C:\APPS\l0gpu16\leelaz --gtp-command "time_settings 1 4 1" -- C:\APPS\l0gpu16\leelaz --gtp-command "time_settings 1 4 1" -k lmgx88-157a

P.S. 2x1080ti

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #217 Posted: Fri Dec 14, 2018 7:47 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
20 game match between #195 and #157
1x1080, 5 min per side per game, no pondering, komi 7.5 , for both matches.

#195 wins 12-8
(average length 248 moves, average time used 243" and 239")

20 game match between #195 and Leela Master GX88
(average length 258 moves, average time used 242" and 237")
#195 wins 12-8


250+ moves... it seems a lot, maybe it would have been less if I'd used -r 10 ?

If someone wants the game, I'll upload them.
For both matches, I used -alternate, so, there are automatic color changes which make the reports look weird (because of the "Colors Exchanged" column in the stats)
The stats :
Attachment:
157_195.jpg
157_195.jpg [ 160.13 KiB | Viewed 688 times ]
Attachment:
195_gx88.jpg
195_gx88.jpg [ 154.48 KiB | Viewed 688 times ]

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #218 Posted: Tue Dec 18, 2018 7:57 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
Rn, with the LZ weights #191 got 2nd, just behind Golaxy, at the recentAI Ryusei 2018.
(see HERE , where there's a link to download the binaries)

20 game match between Rn (with network #195) and LZ0.16 with #195
--device-id 0 --weights C:\...\195.gz --komi 7.5 --const-time 2 for Rn
--noponder --weights C:\...\195.gz for LZ
-time 1s+3s/1 -auto -komi 7.5 for both

twogtp1.5.0, 1x1080, LZ used almost exactly 2sec/move, and Ray used 1.8sec/move
Result 10-10 No duplicate games or errors, All games by resignation, but for two games I discarded because I didn't understand their result :scratch: ... (see image)
Attachment:
hummmm.jpg
hummmm.jpg [ 31.34 KiB | Viewed 491 times ]
Attachment:
Rn is B.zip [9.63 KiB]
Downloaded 11 times
Attachment:
Rn is W.zip [9.71 KiB]
Downloaded 12 times

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #219 Posted: Tue Dec 18, 2018 8:34 am 
Lives with ko
User avatar

Posts: 250
Liked others: 24
Was liked: 72
Can you explain a bit more? You said this is a 20-game match but you only posted the results of 11 games (games #0 --#10). On the other hand, the zip files you provided do contain 20 games (essai2-0...essai2-9, essai3-0...essai3-9).
What is the correspondence between the table you provided and these 20 games?

Also, concerning the two games whose result you don't understand, if you post them we could check manually to see if there is something wrong.

Top
 Profile  
 
Offline
 Post subject: Re: LZ's progression
Post #220 Posted: Tue Dec 18, 2018 9:58 am 
Lives with ko

Posts: 185
Liked others: 2
Was liked: 52
jlt wrote:
Can you explain a bit more?
:D You're right, it's not very clear...
I ran 20 games, 2 of them with dubious results, so I ran 2 more, and the match result (10-10) is with unequivocal games, all won by resignation.
I just discarded the two strange games.
The zip "Rn is B" contains 10 games, with 6 wins for Rn as B, and
The zip "Rn is W" contains 10 games, with 4 wins for Rn as W.

Concerning the curious (discarded) result in the picture, I think
B sees itself as winner (Result_B : B+13.5) and
W sees itself as winner too (Result_W : W+10.5) and
the 3rd column is Result_Referee : ?

I'll run some more games tomorrow, and I'll post the sgf of the dubious ones.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 266 posts ]  Go to page Previous  1 ... 8, 9, 10, 11, 12, 13, 14  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: jaca, Tryss and 4 guests


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