Re: Training an AlphaGo Zero-like algorithm with limited hardware on 7x7 boards

--===============2922614272529100647==
Content-Type: multipart/alternative; boundary="00000000000018d3d3059d1cf422"

--00000000000018d3d3059d1cf422
Content-Type: text/plain; charset="UTF-8"

This is a report after my first day of training my Ataxx network:
https://www.game-ai-forum.org/viewtopic.php?f=24&t=693
Ataxx is played on a 7x7 board. The rules are different, but I expect 7x7
Go would produce similar results. 2k self-play games are more than enough
to produce a huge strength improvement at the beginning.

It would take my system less than one day to generate 285k games on a
single GPU. But speed optimizations are probably not your biggest problem
at the moment.

As I wrote in my previous message, it is important to control the variety
of your self-play game. In my program, I have a function to count the
number of distinct board configurations for each move number of the
self-play games. This way, I can ensure that the same opening is not
replicated too many times.

--00000000000018d3d3059d1cf422
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>This is a report after my first day of training my At=
axx network:<br></div><div><a href=3D"https://www.game-ai-forum.org/viewtop=
ic.php?f=3D24&amp;t=3D693">https://www.game-ai-forum.org/viewtopic.php?f=3D=
24&amp;t=3D693</a></div><div>Ataxx is played on a 7x7 board. The rules are =
different, but I expect 7x7 Go would produce similar results. 2k self-play =
games are more than enough to produce a huge strength improvement at the be=
ginning.</div><div><br></div><div>It would take my system less than one day=
 to generate 285k games on a single GPU. But speed optimizations are probab=
ly not your biggest problem at the moment.</div><div><br></div><div>As I wr=
ote in my previous message, it is important to control the variety of your =
self-play game. In my program, I have a function to count the number of dis=
tinct board configurations for each move number of the self-play games. Thi=
s way, I can ensure that the same opening is not replicated too many times.=
</div></div>

--00000000000018d3d3059d1cf422--

--===============2922614272529100647==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Computer-go mailing list
[email protected]
http://computer-go.org/mailman/listinfo/computer-go

--===============2922614272529100647==--