bug-gnubg
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] Question about the Hint Window Evals


From: Pablo Sangster
Subject: Re: [Bug-gnubg] Question about the Hint Window Evals
Date: Wed, 2 Feb 2005 11:18:54 -0000

Firstly my apologies for remaining anonymous.
My name is Pablo Sangster
Thanks you for your quick response.

Øystein O Johansen wrote:
Hi Mr. X

I must say I don't like replying to anonymous posters and I usually
never reply to anyone that doesn't even state their name. I generally
find it rude and I can't see any good reason for hiding ones identity.

However I will make this one exception from not replying to anonymous
posters, simply because I was the one who made the user interface
changes to fit the pruning net evaluations.

In the Hint window (for checker play) there are buttons marked
Eval ... 0 1 2 3 4
Where can I toggle the pruning on/off  for the 2-ply button?

The evaluation context for these buttons are (unfortunately) hard
coded. It's therefore not possible to turn the pruning off for these
buttons without recompiling. Sorry!

Occasionally when the filters are not wide enough to catch a move
I want evaluated at 2 ply in the hint window, I select the move and
hit the 2 button. But this seems to give me a PRUNED eval.

You can of course press the "..." button and turn off the pruning
neural nets for the evaluations done with the "Eval" button. If I
coded the cache key calculation correctly to take the prune net
setting into account, the evaluation will be redone without use of
the pruning nets.

Yes thank you for this tip. It didn't occur to me.


To smart folks reading: Did I get the cache key calculation right?
If someone can take a look at the cache key code I modified, and
then confirm it's good, I will be thankful.

By the way Mr. X: Are you sure you understand the consept of the
pruning neural nets? It's importatant to understand that the
pruning nets are only used when selecting moves inside the deeper
plies, and then only to prune away bad candidates. Using the pruning
nets does really seldom makes a big difference to an evaluation.

Yes I fully understand how the pruning neural nets work.
However my speciality is backgames and I am finding that there are enough
differences between the prune and full nets to matter.
BTW I think GNU could do with some more training in developing backgames...


This is also why we've choosen to use the pruning nets for the
predefined buttons in hint windows. It's really not much lost, but
the speed increased is really something. That's why it's set to
default on for all deeper ply searches.

-Øystein


-------------------------------------------------------------------
The information contained in this message may be CONFIDENTIAL and is
intended for the addressee only. Any unauthorised use, dissemination
of the information or copying of this message is prohibited. If you
are not the addressee, please notify the sender immediately by return
e-mail and delete this message.
Thank you.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]