KataGo V1.3

[And]

yes, found https://github.com/SabakiHQ/Sabaki/issues/75
except q5go there are no programs with support for rectangular boards?

[go4thewin]

Thanks for the great program! on ogs, wha are the maximum number of visits used by kata-bot? Also, which new neural net is used? This is the best go bot!

[Limeztone]

In a rough estimate how strong are current KataGo (in human ranking) at 1600 visits/move?

Current KataGo:
1.3.1
g170-b20c256x2-s1039565568-d285739972

[lichigo]

Hello,
I heard about katago recently. I am using lizzie with leela zero and is it possible to download a weight(katago file) with lizzie?

[lightvector]

Thanks for the great program! on ogs, wha are the maximum number of visits used by kata-bot? Also, which new neural net is used? This is the best go bot!

It uses 2500 but may go over due to pondering. The going over doesn't add much strength actually (since usually your strength has to do with the moves and mistakes that you're the worst on), it's mainly just to make it always be thinking and responding quickly.

In a rough estimate how strong are current KataGo (in human ranking) at 1600 visits/move?

Umm... 9d+? Human ranks make no sense to use here.

At this point, for small playout numbers, it's close to as strong as Leela Zero's latest networks at equal playouts or visits and possibly somewhat stronger at equal time depending on your hardware setup. For very high playouts, it's somewhat weaker at equal playouts or visits and possibly close to as strong at equal time. The difference between low and high playouts is since KataGo is using a 20 block net and LZ is a 40 block.

KataGo is improving on only 20 blocks and next net should be around 100 Elo better than the one released with v1.3.1 last week (I will post it shortly) - so there's some chance that this amount of gain and/or continued training surpasses LZ without even switching to 40 blocks - although I have that brewing too and KataGo does also plan to increase network size before too long.

[lightvector]

Hello,
I heard about katago recently. I am using lizzie with leela zero and is it possible to download a weight(katago file) with lizzie?

KataGo's releases are here. https://github.com/lightvector/KataGo/releases

The executable version that comes with Lizzie is v1.2 and is out of date. It doesn't support all the features that were released in the last few weeks (japanese rules, stronger handicap play, accurate lead estimation) and is not capable of running the latest stronger neural nets either.

[xela]

The executable version that comes with Lizzie is v1.2 and is out of date.

Just to be clear: v1.3 does work with Lizzie. You need to install Lizzie first (if you haven't already), then install KataGo 1.3, then use the settings menu in Lizzie (or edit the Lizzie config file directly if you prefer) to replace KataGo 1.2 with 1.3.

[Limeztone]

Umm... 9d+? Human ranks make no sense to use here.

At this point, for small playout numbers, ...

Are you saying that KataGo is at 9d+ strength even at small number of visits/move (like 128 visits/move)?

[lightvector]

Umm... 9d+? Human ranks make no sense to use here.

At this point, for small playout numbers, ...

Are you saying that KataGo is at 9d+ strength even at small number of visits/move (like 128 visits/move)?

Probably not that small?

I was referring more to numbers like your 1600. Of course not everyone has a good GPU, but many of the higher-end GPUs from the last couple of years are very easy to get much more than 1600 visits in reasonable amounts of time. Very top-end ones should be able do it in a second or less.

There's a reasonable chance 1600 for LZ/KataGo is large enough to beat any human in the world more than 50% of the time (except for the chance that maybe there is a specific repeatable exploit in some joseki), at relatively quick time controls? But it's not like I've tested that either, I could be way off.

[lightvector]

And here we go. The run is still going. Another week passed so let's have some new neural nets!

https://github.com/lightvector/KataGo/r ... 1.3.1-nets

A stronger 20-block net, and a very strong 15-block net that is almost as strong as ELFv2's 20 block net at equal visits despite being a smaller and faster net. Maybe not exactly as strong? Not sure, my testing so far is noisy. It should be fairly close though.

[xela]

Nice!

If we're going to see a plethora of KataGo nets (which I would welcome), would you like to propose a naming/numbering convention? Perhaps something a little more concise than "I did this analysis using g170-b20c256x2-s1420141824-d350969033"?

[ez4u]

Nice!

If we're going to see a plethora of KataGo nets (which I would welcome), would you like to propose a naming/numbering convention? Perhaps something a little more concise than "I did this analysis using g170-b20c256x2-s1420141824-d350969033"?

@lightvector
This!

[lightvector]

Maybe you could say "g170 20 block s1.42G"? (i.e. 1.42 billion training samples)

s = samples that net was trained on
d = data rows generated at that point in selfplay

So s and d are two indicators of exactly where it is in the run.

At this point, I'm not going to change the convention mid-run, or try to rename the nets and files that have been uploaded, that would be a mess.

If at some point later we started a public distributed run, we could try a different naming convention then.

Edit: Maybe there's something yet better, but I went ahead and updated the all the Github release pages (https://github.com/lightvector/KataGo/releases) suggesting this convention as at least a shorter-hand name.

[go4thewin]

Umm... 9d+? Human ranks make no sense to use here.

At this point, for small playout numbers, ...

Are you saying that KataGo is at 9d+ strength even at small number of visits/move (like 128 visits/move)?

There are some ranking systems that try to figure this out. According to those rankings, katago 1.2 at 100 playouts was at least 6 Dan.
https://github.com/breakwa11/GoAIRatings Breakwa's rankings estimate leela zero 157 net at 100 playouts ranked at 6d human. CGOS has katago 1.2 at 100 playouts ranked higher (Bayle's ELO) than a leela zero 157 version (9006c708) at 200 playouts. http://www.yss-aya.com/cgos/19x19/cross ... _v100.html
These rankings also estimate that katago 1.3.1 at 800 playouts would rank higher than any human player in the world in even games, though I can't believe that. These are kind of just guesses.

[jlt]

When playing handicap games against Katago, the initial board positions are not always the same, but look like this:

$$W
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . X . . . . . . . . . . . . . |
$$ | . . X , . . . . . , . . . X . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X , . X . . . , . . . X . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------

Click Here To Show Diagram Code

[go]$$W
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . X . . . . . . . . . . . . . |
$$ | . . X , . . . . . , . . . X . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X , . X . . . , . . . X . , X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Is there a way to choose myself where to place the handicap stones?