It is currently Tue Feb 19, 2019 10:47 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 26 posts ]  Go to page Previous  1, 2
Author Message
Offline
 Post subject: Re: so basically leela zero has made 0 progress in the last
Post #21 Posted: Sat Nov 24, 2018 9:31 am 
Lives with ko

Posts: 127
Liked others: 19
Was liked: 40
Rank: 2d
Relevant discussion elsewhere:
https://www.reddit.com/r/cbaduk/comments/9zv4v1/leelazero_40b_networks_are_overfit/


This post by bernds was liked by: ez4u
Top
 Profile  
 
Offline
 Post subject: Re: so basically leela zero has made 0 progress in the last
Post #22 Posted: Sat Nov 24, 2018 10:58 pm 
Oza
User avatar

Posts: 2095
Location: Tokyo, Japan
Liked others: 1917
Was liked: 1188
Rank: Jp 6 dan
KGS: ez4u
bernds wrote:

This is a very interesting report! The follow up posting on the LZ site is here: https://github.com/gcp/leela-zero/issues/2044

_________________
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21


This post by ez4u was liked by 2 people: dfan, Gomoto
Top
 Profile  
 
Offline
 Post subject: Re: so basically leela zero has made 0 progress in the last
Post #23 Posted: Sun Nov 25, 2018 1:53 pm 
Lives with ko

Posts: 248
Liked others: 208
Was liked: 128
ez4u wrote:
The follow up posting on the LZ site is here: https://github.com/gcp/leela-zero/issues/2044


To summarize the github thread: the claim that starting with pre-set openings somehow makes 40 block LZ weaker than 15 block is wrong.

The error was due to using incorrect setup params, which somehow seem to affect pre-set games but not regular games.

Phew, the world is sane again :-)

_________________
Sorin - 361points.com


This post by sorin was liked by 2 people: ez4u, Gomoto
Top
 Profile  
 
Offline
 Post subject: Re: so basically leela zero has made 0 progress in the last
Post #24 Posted: Fri Nov 30, 2018 6:25 am 
Lives with ko

Posts: 199
Liked others: 2
Was liked: 53
The networks do get better, even at short time parity (~1 sec/move) , see HERE

Top
 Profile  
 
Offline
 Post subject: Re: so basically leela zero has made 0 progress in the last
Post #25 Posted: Mon Dec 24, 2018 8:23 am 
Dies in gote

Posts: 39
Liked others: 0
Was liked: 7
OGS: roy7
Universal go server handle: roy7
bernds wrote:
I was curious, so I was trying to run something similar yesterday. While I got a similar result, the games all looked identical up to move 40 or so.


I think I saw a comment in the github by someone else who ran comparisons that one of the tools people sometimes use to run these head to head matches (twogtp?) doesn't restart the engine between games, it just uses the existing engine in memory and sets up the new board/etc and begins playing the new game. The issue with this at the time (perhaps a recent pull request addressed it) is that the NN eval cache doesn't flush just because you cleared the board. Thus you'll have exactly the same results of every early NN eval each game you play. (Normally symmetry causes some small random noise, but once cached the value never changes.)

Top
 Profile  
 
Offline
 Post subject: Re: so basically leela zero has made 0 progress in the last
Post #26 Posted: Sat Jan 19, 2019 12:04 pm 
Dies in gote

Posts: 53
Liked others: 0
Was liked: 3
So basically I was wrong,

make that half a year zero net effective progress ...

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 26 posts ]  Go to page Previous  1, 2

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: cobrahc, Google [Bot] and 6 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