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

--===============6713588426182963447==
Content-Type: multipart/alternative; boundary="000000000000263cd7059d0d255d"

--000000000000263cd7059d0d255d
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Yes, using komi would help a lot. Still, I feel that something else must be
wrong, because winning 100% of the games as Black without komi should be
very easy on 7x7.

I have not written anything about what I did with Crazy Stone. But my
experiments and ideas were really very similar to what David Wu did:
https://blog.janestreet.com/accelerating-self-play-learning-in-go/

To clarify what I wrote in my previous message: "strong from scratch in a
single day" was for 7x7. I like testing new ideas with small networks on
small boards, because training is very fast, and what works on small boards
with small networks usually also works on large boards with big networks.

R=C3=A9mi

On Sun, Jan 26, 2020 at 12:30 AM cody2007 <[email protected]> wrote:

> Hi R=C3=A9mi,
>
> Thanks for your comments! I am not using any komi and had not given much
> thought to it. Although, I suppose by having black win most games, I'm
> depriving the network of its only learning signal. I will have to try wit=
h
> an appropriately set komi next...
>
> >When I started to develop the Zero version of Crazy Stone, I spend a lot
> of time optimizing my method on a single (V100) GPU
> Any chance you've written about it somewhere? I'd be interested to learn
> more but wasn't able to find anything on the Crazy Stone website.
>
> Thanks,
> Cody
>
> =E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90 Original =
Message =E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90
> On Saturday, January 25, 2020 5:49 PM, R=C3=A9mi Coulom <remi.coulom@gmai=
l.com>
> wrote:
>
> Hi,
>
> Thanks for sharing your experiments.
>
> Your match results are strange. Did you use a komi? You should use a komi
> of 9:
> https://senseis.xmp.net/?7x7
>
> The final strength of your network looks surprisingly weak. When I starte=
d
> to develop the Zero version of Crazy Stone, I spend a lot of time
> optimizing my method on a single (V100) GPU. I could train a strong netwo=
rk
> from scratch in a single day. Using a wrong komi might have hurt you. Als=
o,
> on such a small board, it is not so easy to make sure that the self-play
> games have enough variety. You'd have to find many balanced random initia=
l
> positions in order to avoid replicating the same game again and again.
>
> R=C3=A9mi
>
>
>

--000000000000263cd7059d0d255d
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr"><div>Yes, using komi would help a lot. St=
ill, I feel that something else must be wrong, because winning 100% of the =
games as Black without komi should be very easy on 7x7.</div><div><br></div=
><div>I have not written anything about what I did with Crazy Stone. But my=
 experiments and ideas were really very similar to what David Wu did:</div>=
<div><a href=3D"https://blog.janestreet.com/accelerating-self-play-learning=
-in-go/">https://blog.janestreet.com/accelerating-self-play-learning-in-go/=
</a></div><div><br></div><div>To clarify what I wrote in my previous messag=
e: &quot;strong from scratch in a single day&quot; was for 7x7. I like test=
ing new ideas with small networks on small boards, because training is very=
 fast, and what works on small boards with small networks usually also work=
s on large boards with big networks.</div><div><br></div><div>R=C3=A9mi<br>=
</div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_=
attr">On Sun, Jan 26, 2020 at 12:30 AM cody2007 &lt;<a href=3D"mailto:cody2=
[email protected]">[email protected]</a>&gt; wrote:<br></div><blockq=
uote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1p=
x solid rgb(204,204,204);padding-left:1ex"><div>Hi R=C3=A9mi,<br></div><div=
><br></div><div>Thanks for your comments! I am not using any komi and had n=
ot given much thought to it. Although, I suppose by having black win most g=
ames, I&#39;m depriving the network of its only learning signal. I will hav=
e to try with an appropriately set komi next...<br></div><div><br></div><di=
v>&gt;When I started to develop the Zero version of Crazy Stone, I spend a =
lot of time optimizing my method on a single (V100) GPU<br></div><div>Any c=
hance you&#39;ve written about it somewhere? I&#39;d be interested to learn=
 more but wasn&#39;t able to find anything on the Crazy Stone website.<br><=
/div><div><br></div><div>Thanks,<br></div><div>Cody<br></div><div><br></div=
><div>=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90 Origi=
nal Message =E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=E2=80=90=
<br></div><div> On Saturday, January 25, 2020 5:49 PM, R=C3=A9mi Coulom &lt=
;<a href=3D"mailto:[email protected]" target=3D"_blank">remi.coulom@gma=
il.com</a>&gt; wrote:<br></div><div> <br></div><blockquote type=3D"cite"><d=
iv dir=3D"ltr"><div>Hi,<br></div><div><br></div><div>Thanks for sharing you=
r experiments.<br></div><div><br></div><div>Your match results are strange.=
 Did you use a komi? You should use a komi of 9:<br></div><div><a href=3D"h=
ttps://senseis.xmp.net/?7x7" target=3D"_blank">https://senseis.xmp.net/?7x7=
</a><br></div><div><br></div><div>The final strength of your network looks =
surprisingly weak. When I started to develop the Zero version of Crazy Ston=
e, I spend a lot of time optimizing my method on a single (V100) GPU. I cou=
ld train a strong network from scratch in a single day. Using a wrong komi =
might have hurt you. Also, on such a small board, it is not so easy to make=
 sure that the self-play games have enough variety. You&#39;d have to find =
many balanced random initial positions in order to avoid replicating the sa=
me game again and again.<br></div><div><br></div><div>R=C3=A9mi<br></div></=
div></blockquote><div><br></div></blockquote></div></div>

--000000000000263cd7059d0d255d--

--===============6713588426182963447==
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

--===============6713588426182963447==--