It is currently Sat May 10, 2025 11:14 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 67 posts ]  Go to page Previous  1, 2, 3, 4  Next
Author Message
Offline
 Post subject: Re: Crazystone
Post #21 Posted: Fri May 13, 2016 12:37 am 
Lives with ko

Posts: 170
Liked others: 32
Was liked: 119
Rank: KGS 4 kyu
ez4u wrote:
Remi, Why did you play it so quickly? On that level of hardware, what difference would 10 seconds per move or 20 seconds per move have made?

I was calibrating weak levels for the commercial release. I did not expect it would reach 7d. Next week I'll connect Crazy Stone on a powerful server. I believe there is a chance that it may reach 8d. We'll see.

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #22 Posted: Fri May 13, 2016 1:27 am 
Lives in sente

Posts: 727
Liked others: 44
Was liked: 218
GD Posts: 10
Rémi wrote:
ez4u wrote:
Remi, Why did you play it so quickly? On that level of hardware, what difference would 10 seconds per move or 20 seconds per move have made?

I was calibrating weak levels for the commercial release. I did not expect it would reach 7d. Next week I'll connect Crazy Stone on a powerful server. I believe there is a chance that it may reach 8d. We'll see.

On KGS? I'll stay awake for the whole week!!!

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #23 Posted: Fri May 13, 2016 2:19 am 
Dies with sente

Posts: 93
Liked others: 26
Was liked: 13
Rank: Total beginner
Universal go server handle: Satorian
I've got a desktop running an i7 4790k at 4 GHz, and a laptop running an i7 6700HQ. Looking forward to how well they do. :)

The desktop is also equipped with a GTX 980Ti, while the laptop has got a 960M, both with current CUDA installations. Wouldn't mind seeing a GPU-implementation of CS' DCNN at some point.

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #24 Posted: Sat May 14, 2016 2:08 am 
Beginner

Posts: 1
Liked others: 0
Was liked: 0
Three quick questions @Remi:

1. Do you get significant improvement from your neural networks in 13x13 and 9x9? If so, is it correct to assume that the training of the small board variants was completely separate from the training of the standard size net?

2. Will the play of the weak pure DCNN levels adjust to komi?

3. Does CSDL have an analogue to the AlphaGo value network or just to the policy network at the moment?

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #25 Posted: Sat May 14, 2016 10:32 am 
Lives in gote

Posts: 677
Liked others: 6
Was liked: 31
KGS: 2d
I will wait and see how strong CS will be on weaker machines. I have a laptop with duo core, 2,4 GHz and 4GB RAM and I am 2d-3d strong, so I have some doubt if CS can be at least one full stone stronger than me (I want to use it for analysis and as a teacher).

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #26 Posted: Sat May 14, 2016 10:39 am 
Dies with sente

Posts: 93
Liked others: 26
Was liked: 13
Rank: Total beginner
Universal go server handle: Satorian
Pippen wrote:
I will wait and see how strong CS will be on weaker machines. I have a laptop with duo core, 2,4 GHz and 4GB RAM and I am 2d-3d strong, so I have some doubt if CS can be at least one full stone stronger than me (I want to use it for analysis and as a teacher).


You can set fixed time limits per move or per game IIRC for whole game analysis. So, if you don't need that 7-8d strength live while playing, your configuration should be more than enough to be able to analyze games at that level when you give CS 2-3h of time.

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #27 Posted: Sat May 14, 2016 12:02 pm 
Lives in sente
User avatar

Posts: 1326
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
Here comes Rémi's kind answer to a similar question of mine, related to "Analyse Current Game" = analyse the current position = show the next move's candidates.
This feature (for a single move) is not the same as "Analyse Game Record ..." (for every move contained in a SGF's main line), which is mentioned in the posting above, but I think that Rémi's core message will apply to that as well.

"The analysis will be of the same quality for the same number of playouts, whatever the hardware. Of course, the same number of playouts will be faster on a more powerful machine."

So you will be able to get a high-level analysis done be CS, if time is no limiting factor for you.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #28 Posted: Sat May 14, 2016 2:44 pm 
Lives in gote

Posts: 677
Liked others: 6
Was liked: 31
KGS: 2d
Cassandra wrote:
"The analysis will be of the same quality for the same number of playouts, whatever the hardware. Of course, the same number of playouts will be faster on a more powerful machine."

So you will be able to get a high-level analysis done be CS, if time is no limiting factor for you.


Thx for the info. If Remi can confirm that it'd be a huge argument to buy CS.

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #29 Posted: Sat May 14, 2016 6:10 pm 
Dies in gote

Posts: 51
Liked others: 3
Was liked: 5
Pippen wrote:
Cassandra wrote:
"The analysis will be of the same quality for the same number of playouts, whatever the hardware. Of course, the same number of playouts will be faster on a more powerful machine."

So you will be able to get a high-level analysis done be CS, if time is no limiting factor for you.


Thx for the info. If Remi can confirm that it'd be a huge argument to buy CS.

You can check out this thread where Remi seemed to have confirmed it, but you may want another confirmation since this is another Crazy Stone version, and I am not sure if there is a timed mode for the new version:
http://www.lifein19x19.com/forum/viewto ... 1&start=40
Rémi wrote:
In terms of speed, Crazy stone does about 250 playouts / second / core on the transformer prime, so 1000 playouts / second on 4 cores. This compares to about 2000 playouts / second /core on a PC. Except for speed, the program is otherwise identical.

Rémi

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #30 Posted: Sun May 15, 2016 1:32 am 
Lives in sente
User avatar

Posts: 1326
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
pragmaticleas wrote:
You can check out this thread where Remi seemed to have confirmed it, but you may want another confirmation since this is another Crazy Stone version, and I am not sure if there is a timed mode for the new version:

CS 2012 has it, CS 2013 has it. CS 2016 screenshots look like the same.
Analysis features are explicitely mentioned in the Unbalance announcement.

Why should they make a step back ?

In my opinion, you will use the "Analyse Game Record ..." feature to identify the "hot spots" of a game seen by CS (e.g. where CS's favourite move has a much better winning percentage than the really played one).
CS 2012 has about 30.000 playouts per move with the setting "depth = high". (Lasted about 20 to 30 minutes in total on my low-level laptop).
I did not try the "Total time" option myself, but the maximum time that can be chosen is 8 hours !

"Analyse Current Game" would be used to get an idea of the moves CS has thought about (additionally maybe the histogram) in a special position, and to try out variations.
This analysis for one special move will last until you stop it (e.g. by going to the next move played).

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #31 Posted: Tue May 17, 2016 2:17 pm 
Dies in gote

Posts: 26
Liked others: 4
Was liked: 9
Crazystone continues to be a really impressive bot. I just bought my copy last night and I was up to wee hours analyzing some games, watching it play itself, and relaying a game for a KGS buddy.

I wish I could set it up for kgsgtp.

I have to say that I still really like it's style, it reminds me a lot of old go games from China (1300s-1700s). (I could probably ramble on about this subject for some time.) While it seems to value influence much more than they did, it has a very similar sense of trying to discoordinate opponent stones. (I could probably ramble on this subject for some time as well.) I had SGFs of two old games transcribed from "The Celestial Arsenal" and gave CS a chance to analyze them. It seems CS appreciated the style a lot, but I am not sure I understand the abstract number given by analysis mode for how much CS likes its own suggested move relative to the move made.

Analysis is a little clunky if you want to play out a tree, but since you can just turn CS on with no time limit you can see what it's thinking anyway by playing against it, or having it play itself for a few moves.

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #32 Posted: Tue May 17, 2016 2:29 pm 
Gosei

Posts: 1596
Liked others: 891
Was liked: 533
Rank: AGA 2k Fox 3d
GD Posts: 61
KGS: dfan
erislover wrote:
I am not sure I understand the abstract number given by analysis mode for how much CS likes its own suggested move relative to the move made.

It's how much it thinks the probability of winning the game would have gone up if the suggested move had been played rather than the actual move. Unfortunately,
  • It's given to 6 digits, which is hard to read, has much more precision than is actually relevant, and implies much more precision than actually exists.
  • It's given as a raw number, rather than a percentage, which means every number starts with a "0.0" you have to mentally strip off.
As an example, if the Delta it gives is 0.023193, this means that it thinks that the player's chance of winning would have been 2.3% higher if he had played the suggested move instead.

(I think that the "probability of winning" is really what fraction of playouts were wins, but that's kind of a detail. It would mean that all the probabilities are closer to 50% than you'd expect.)

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #33 Posted: Tue May 17, 2016 2:35 pm 
Dies in gote

Posts: 26
Liked others: 4
Was liked: 9
dfan wrote:
It's how much it thinks the probability of winning the game would have gone up if the suggested move had been played rather than the actual move.
That's what I thought, thanks!

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #34 Posted: Tue May 17, 2016 4:25 pm 
Lives in gote

Posts: 677
Liked others: 6
Was liked: 31
KGS: 2d
If I buy CS and install it and then archive it on DVD, can I use the DVd copy to reinstall it on another computer or if my original gets deleted? I don't wanna spend $80 for a one-and-done-copy. If possible do I have to be connected to the Internet for the second install again?

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #35 Posted: Wed May 18, 2016 2:25 am 
Dies in gote

Posts: 38
Liked others: 4
Was liked: 20
Pippen wrote:
If I buy CS and install it and then archive it on DVD, can I use the DVd copy to reinstall it on another computer or if my original gets deleted? I don't wanna spend $80 for a one-and-done-copy. If possible do I have to be connected to the Internet for the second install again?


Hmm, there was some talk of a Steam version of CrazyStone, which I think would let you use it on any machine by logging into your Steam account. I wonder if that's happening though.

https://steamcommunity.com/sharedfiles/ ... /389714638

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #36 Posted: Wed May 18, 2016 5:58 am 
Dies with sente

Posts: 93
Liked others: 26
Was liked: 13
Rank: Total beginner
Universal go server handle: Satorian
Rémi wrote:
The top levels are in fact defined as time per move, so they will be stronger with more powerful hardware.


If we measure playout speed in analysis mode, would there be a way to look up po/s versus approximate dan rating at each setting? Do you have a table like that?

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #37 Posted: Wed May 18, 2016 8:16 am 
Lives in sente

Posts: 1045
Liked others: 0
Was liked: 182
Pippen wrote:
If I buy CS and install it and then archive it on DVD, can I use the DVd copy to reinstall it on another computer or if my original gets deleted? I don't wanna spend $80 for a one-and-done-copy. If possible do I have to be connected to the Internet for the second install again?


Those are the sorts of questions that a vendor of software SHOULD make clear. Whether the license is for one INSTALL, for use on ONE MACHINE (ever) regardless of how many times needs to be reinstalled, or for ONE MACHINE (at a time), meaning you can move it to a new machine when your old one is being replaced, or even a "site license" meaning can be on more than one of YOUR machines << at least of the go programs being sold is on that basis >>

Note that while you are referring to the cost of this software, the "free software" movement did not originally envision free as in free beer downloads but fairer license terms and low cost*, essentially comparable to a book (but note since academics, probably thinking in terms of a college textbook, not a paperback romance novel).

IMHO --- $80 would be steep for "single install" but quite reasonable for "use on one machine" (at a time) and a steal if "site license".

* Remember, this was in the days before much broadband internet and not just a few dominant OS's in use. Which is why in theory can be charged for the FREE software (source code --- mailed to you on standard medium for the accepted fee for that sort of service) and anybody could compile and link executables for running under whatever OS charging whatever they wanted for doing that for you with the idea being that if they were too greedy somebody would step up and offer that service for less.

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #38 Posted: Wed May 18, 2016 10:09 am 
Lives in gote

Posts: 677
Liked others: 6
Was liked: 31
KGS: 2d
@Mike: I agree, but Unbalance is silence about all that. You basically buy the cat in the bag. It's only written that your computer has to have internet access for the first install. I'll wait a bit anyway to see what users say about it, so we'll see. Maybe soon other programs of that kind will flood the market (Zen? DolBaram?...)

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #39 Posted: Wed May 18, 2016 10:17 am 
Dies in gote

Posts: 23
Liked others: 9
Was liked: 13
I bought the 2013 version in January 2015. Later that year I sold my PC and bought a new one. I just installed the program, put in the serial number on the first startup just like before and then it just worked, no problems at all.

Top
 Profile  
 
Offline
 Post subject: Re: Crazystone
Post #40 Posted: Wed May 18, 2016 10:18 am 
Oza
User avatar

Posts: 2777
Location: Seattle, WA
Liked others: 251
Was liked: 549
KGS: oren
Tygem: oren740, orenl
IGS: oren
Wbaduk: oren
Pippen wrote:
@Mike: I agree, but Unbalance is silence about all that. You basically buy the cat in the bag. It's only written that your computer has to have internet access for the first install. I'll wait a bit anyway to see what users say about it, so we'll see. Maybe soon other programs of that kind will flood the market (Zen? DolBaram?...)


Zen announced a June release.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 67 posts ]  Go to page Previous  1, 2, 3, 4  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: Bing [Bot] 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