Moses-support Digest, Vol 94, Issue 10

Send Moses-support mailing list submissions to
moses-support@mit.edu

To subscribe or unsubscribe via the World Wide Web, visit
http://mailman.mit.edu/mailman/listinfo/moses-support
or, via email, send a message with subject or body 'help' to
moses-support-request@mit.edu

You can reach the person managing the list at
moses-support-owner@mit.edu

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Moses-support digest..."


Today's Topics:

1. kbmira does not work (Toshiaki Nakazawa)
2. Re: kbmira does not work (Barry Haddow)


----------------------------------------------------------------------

Message: 1
Date: Fri, 08 Aug 2014 17:04:35 +0900
From: Toshiaki Nakazawa <nakazawa@pa.jst.jp>
Subject: [Moses-support] kbmira does not work
To: moses-support@mit.edu
Message-ID: <m2ppgb9zl8.wl%nakazawa@pa.jst.jp>
Content-Type: text/plain; charset=US-ASCII

Hi all,

I'm now trying the latest version of Moses (git pull and bjam -a
today). When I use kbmira with some data, it dies after the first
update.

$ mosesdecoder/bin/kbmira --scfile run0.scores.dat --ffile run0.features.dat
kbmira with c=0.01 decay=0.999 no_shuffle=0
Initialising random seed from system clock
.................Initial BLEU = 0.135127
1013/1790 updates, avg loss = 1.67374, BLEU = 0.166537
zsh: segmentation fault nice -19 /Users/nakazawa/work/MT/tool/mosesdecoder/bin/kbmira --scfile -o

The data are
http://lotus.kuee.kyoto-u.ac.jp/~nakazawa/run0.scores.dat
http://lotus.kuee.kyoto-u.ac.jp/~nakazawa/run0.features.dat

$ g++ --version
Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/usr/include/c++/4.2.1
Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn)
Target: x86_64-apple-darwin13.3.0
Thread model: posix

Does someone have any ideas?

Thanks,

Toshiaki Nakazawa


------------------------------

Message: 2
Date: Fri, 08 Aug 2014 11:11:07 +0100
From: Barry Haddow <bhaddow@staffmail.ed.ac.uk>
Subject: Re: [Moses-support] kbmira does not work
To: Toshiaki Nakazawa <nakazawa@pa.jst.jp>, moses-support@mit.edu
Message-ID: <53E4A23B.9040403@staffmail.ed.ac.uk>
Content-Type: text/plain; charset="iso-8859-1"

Hi Toshiaki

Thanks for the clear bug report.

If you supply kbmira with an initial weights file, like the one
attached, using the --dense-init argument, then it does not seg fault.
This is supplied automatically if you use kbmira from within mert-moses.pl

I will try to fix kbmira so that it at least gives an error message, or
supplies a sensible default, when no weights file is given,

cheers - Barry

On 08/08/14 09:04, Toshiaki Nakazawa wrote:
> Hi all,
>
> I'm now trying the latest version of Moses (git pull and bjam -a
> today). When I use kbmira with some data, it dies after the first
> update.
>
> $ mosesdecoder/bin/kbmira --scfile run0.scores.dat --ffile run0.features.dat
> kbmira with c=0.01 decay=0.999 no_shuffle=0
> Initialising random seed from system clock
> .................Initial BLEU = 0.135127
> 1013/1790 updates, avg loss = 1.67374, BLEU = 0.166537
> zsh: segmentation fault nice -19 /Users/nakazawa/work/MT/tool/mosesdecoder/bin/kbmira --scfile -o
>
> The data are
> http://lotus.kuee.kyoto-u.ac.jp/~nakazawa/run0.scores.dat
> http://lotus.kuee.kyoto-u.ac.jp/~nakazawa/run0.features.dat
>
> $ g++ --version
> Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/usr/include/c++/4.2.1
> Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn)
> Target: x86_64-apple-darwin13.3.0
> Thread model: posix
>
> Does someone have any ideas?
>
> Thanks,
>
> Toshiaki Nakazawa
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>


--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

-------------- next part --------------
f0= 0.1
f1= 0.1
f2= 0.1
f3= 0.1
f4= 0.1
f5= 0.1
f6= 0.1
f7= 0.1
f8= 0.1
f9= 0.1
f10= 0.1
f11= 0.1
f12= 0.1
f13= 0.1
f14= 0.1
f15= 0.1
f16= 0.1
f17= 0.1
f18= 0.1
f19= 0.1
f20= 0.1
f21= 0.1
f22= 0.1
f23= 0.1
f24= 0.1
f25= 0.1
f26= 0.1
f27= 0.1
f28= 0.1
f29= 0.1
f30= 0.1
f31= 0.1
f32= 0.1
f33= 0.1
f34= 0.1
f35= 0.1
f36= 0.1
f37= 0.1
f38= 0.1
f39= 0.1
f40= 0.1
f41= 0.1

------------------------------

_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support


End of Moses-support Digest, Vol 94, Issue 10
*********************************************

Tweet