Home » Uncategories » Moses-support Digest, Vol 89, Issue 75

Moses-support Digest, Vol 89, Issue 75

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. factored training data question (Viktor Pless)
2. Re: windows cygwin moses build (jfrm.maurel@gmail.com)
3. Evaluation of rule based system (Rajkiran Rajkumar)
4. <factor> in train-model.perl's -lm parameter (Benjamin K?rner)
5. Re: Evaluation of rule based system (Philipp Koehn)
6. Re: <factor> in train-model.perl's -lm parameter (Philipp Koehn)


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

Message: 1
Date: Fri, 28 Mar 2014 23:40:38 +0100
From: Viktor Pless <viktor.pless@gmail.com>
Subject: [Moses-support] factored training data question
To: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAD3iPeFT1Q68qOqjSiAVmyJ+qS58V2QCpxd=TyQbB0qYu1Dxgw@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

Hi everyone,
How can I produce factored training data like below?

"You will have to provide training data in the format
word0factor0|word0factor1|word0factor2
word1factor0|word1factor1|word1factor2 ..."

Could you please tell me the what program produces this format? Neither
MXPOST nor treetagger does it for me.

Thank you in advance.
Viktor
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20140328/b3dea345/attachment-0001.htm

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

Message: 2
Date: Sat, 29 Mar 2014 09:26:55 +0100
From: "jfrm.maurel@gmail.com" <jfrm.maurel@gmail.com>
Subject: Re: [Moses-support] windows cygwin moses build
To: moses-support@mit.edu
Cc: Hieu Hoang <hieu.hoang@ed.ac.uk>
Message-ID: <533683CF.5070406@gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

Le 24/03/2014 21:12, Hieu Hoang a ?crit :

Hi Hieu,

> There seems to be a problem using the boost library

I have been told by boost.context people that this library is not
available under cygwin:
http://article.gmane.org/gmane.comp.lib.boost.user/81431

Could you please tell me if boost.context is mandatory in order to build
Moses ?

Regards

--
Jean-Fran?ois MAUREL
PIMECA
http://www.pimeca.com



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

Message: 3
Date: Sat, 29 Mar 2014 14:07:15 +0530
From: Rajkiran Rajkumar <rajkiran2507@gmail.com>
Subject: [Moses-support] Evaluation of rule based system
To: moses-support@mit.edu
Message-ID:
<CAHMStsKJ8hnTCWbJ9o4-F-FgSxYVtdxMBxw_J8hvHe6DUL3aEw@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

Hi,
I am trying to evaluate a rule based (hierarchical) MT system from Tamil to
English. The translation works, but I am not able to run multi-bleu.perl on
it.
The error I am getting is,
*ERROR: could not find reference file corpus.bcn.test.true.en at
/home/rajkiran/mosesdecoder/scripts/generic/multi-bleu.perl line 28.*
The said file exists in the location. I've checked it.
Where am I going wrong?
Apologies if this is a repeat question. I couldn't find threads relevant to
this.

Regards,
Rajkiran
College of Engineering, Guindy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20140329/4de9d284/attachment-0001.htm

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

Message: 4
Date: Sat, 29 Mar 2014 11:46:07 +0100
From: Benjamin K?rner <b.koerner@stud.uni-heidelberg.de>
Subject: [Moses-support] <factor> in train-model.perl's -lm parameter
To: <moses-support@mit.edu>
Message-ID: <002501cf4b3c$17c54940$474fdbc0$@stud.uni-heidelberg.de>
Content-Type: text/plain; charset="iso-8859-1"

Hi guys,



I'm wondering what those factor codes mean. 0 is for an unfactored model,
right? What about 1-9 then?



Thanks,



Benjamin



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20140329/6e535a37/attachment-0001.htm

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

Message: 5
Date: Sat, 29 Mar 2014 11:35:37 +0000
From: Philipp Koehn <pkoehn@inf.ed.ac.uk>
Subject: Re: [Moses-support] Evaluation of rule based system
To: Rajkiran Rajkumar <rajkiran2507@gmail.com>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAAFADDC4iU_Gh1NdGUCe2b_m+BSfwYkmv7T26zaePtZHy8khjA@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

Hi,

I could not even easily say in which directory the script is looking
for the file, so please specify a filename with full path, so it is not
confused.

-phi

On Sat, Mar 29, 2014 at 8:37 AM, Rajkiran Rajkumar
<rajkiran2507@gmail.com> wrote:
> Hi,
> I am trying to evaluate a rule based (hierarchical) MT system from Tamil to
> English. The translation works, but I am not able to run multi-bleu.perl on
> it.
> The error I am getting is,
> ERROR: could not find reference file corpus.bcn.test.true.en at
> /home/rajkiran/mosesdecoder/scripts/generic/multi-bleu.perl line 28.
> The said file exists in the location. I've checked it.
> Where am I going wrong?
> Apologies if this is a repeat question. I couldn't find threads relevant to
> this.
>
> Regards,
> Rajkiran
> College of Engineering, Guindy
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>


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

Message: 6
Date: Sat, 29 Mar 2014 11:37:51 +0000
From: Philipp Koehn <pkoehn@inf.ed.ac.uk>
Subject: Re: [Moses-support] <factor> in train-model.perl's -lm
parameter
To: Benjamin K?rner <b.koerner@stud.uni-heidelberg.de>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAAFADDBcdapFHhAw1Cq8UJKe8GF5dJMSui90zEJZFLoq--_20w@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

Hi,

model training and decoding is generally unaware of what the factors are,
but factor "0" is considered the surface factor. If you build a model that
generates additional output factors (POS tags, word classes, whathaveyou),
then you can build language models for these as well.

-phi


On Sat, Mar 29, 2014 at 10:46 AM, Benjamin K?rner
<b.koerner@stud.uni-heidelberg.de> wrote:
> Hi guys,
>
>
>
> I'm wondering what those factor codes mean. 0 is for an unfactored model,
> right? What about 1-9 then?
>
>
>
> Thanks,
>
>
>
> Benjamin
>
>
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>



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

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


End of Moses-support Digest, Vol 89, Issue 75
*********************************************

0 Response to "Moses-support Digest, Vol 89, Issue 75"

Post a Comment

Subscribe to: Post Comments (Atom)