Moses-support Digest, Vol 101, Issue 50

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. Re: -print-alignment-info option (Hieu Hoang)
2. Missing processPhraseTable in Release 3.0
(Ertu?rul YILMAZ (B?LGEM))
3. Can't train model on factored corpus with 0-0 translation
factors? (Stanislav Ku??k)
4. Support for Backoff Rule Tables? (Jeremy Gwinnup)
5. Re: Missing processPhraseTable in Release 3.0
(Muhammad Danial Raza)


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

Message: 1
Date: Mon, 16 Mar 2015 16:58:46 +0000
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] -print-alignment-info option
To: Tom Hoar <tahoar@precisiontranslationtools.com>
Cc: moses-support <moses-support@mit.edu>
Message-ID:
<CAEKMkbj2T2-Y7tSTO5bYvjeff4+fVdscuFVgZ=6_qDcAkCEWtA@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

You're saying it gives a different translation depending on whether
-print-alignment-info is used or not? That would be a very strange
behaviour. Which phrase-table are you using?

If you can share your model files, i'll take a look at it.

Hieu Hoang
Research Associate (until March 2015)
University of Edinburgh
http://www.hoang.co.uk/hieu

On 16 March 2015 at 02:53, Tom Hoar <tahoar@precisiontranslationtools.com>
wrote:

> This is my first time using the '-print-alignment-info' option and I'm
> getting some unexpected results.
>
> In a set of five sentences, when using no option or
> '-report-segmentation' or '-report-segmentation-enriched', all five
> sentence translations are output with respective alignment info. With
> the '-print-alignment-info' option, two sentence translations are output
> as zero-length strings.
>
> Is this what is meant when the help prompt says: "Word-to-word
> alignments are taken from the phrase table if any"?
>
> When no word-to-word alignment info is in the phrase table, I would
> expect the translated output followed by " ||| " and then nothing.
>
> Thanks.
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150316/074149da/attachment-0001.htm

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

Message: 2
Date: Tue, 17 Mar 2015 15:32:21 +0200 (EET)
From: Ertu?rul YILMAZ (B?LGEM) <yilmaz.ertugrul@tubitak.gov.tr>
Subject: [Moses-support] Missing processPhraseTable in Release 3.0
To: moses-support <moses-support@mit.edu>
Message-ID:
<433283511.51728832.1426599141886.JavaMail.zimbra@tubitak.gov.tr>
Content-Type: text/plain; charset="utf-8"

Hi All,

I am upgrading our systems to the newest moses release (3.0), downloaded and compiled the sources as shown below

git clone https://github.com/moses-smt/mosesdecoder.git --branch RELEASE-3.0 --single-branch RELEASE-3.0
/usr/bin/bjam --with-boost=/home/ertugrul/tools/boost_1_55_0 -j8

All completed with success. But I am missing the /bin/processPhraseTable executable in my executables. Has that executable been replaced by something else, or have I missed something along the way?

Can someone provide information on this, pleaase?

Thanks much

----------------------------------------------------------
Ertu?rul YILMAZ


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150317/9e4722ae/attachment-0001.htm

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

Message: 3
Date: Tue, 17 Mar 2015 13:32:49 +0000 (UTC)
From: Stanislav Ku??k <standa.kurik@gmail.com>
Subject: [Moses-support] Can't train model on factored corpus with 0-0
translation factors?
To: moses-support@mit.edu
Message-ID: <loom.20150317T140232-923@post.gmane.org>
Content-Type: text/plain; charset=utf-8

Hello,

it looks to me that --translation-factors=0-0 cannot be used as a
special
case of factored corpus training. I am able to successfully train a
model
with --translation-factors=0-0,2. However, when I change it to
--translation-factors=0-0 (and that's the only change), the training
finishes successfully, but

1) The phrase table doesn't look right:

after|after|IN the|the|DT ||| po|po|RR zform?tov?n?|zform?tov?n?|NN ...
|

(shortened). There are all factors present, although only the factor 0
should be on both sides, if I am not mistaken.

2) The model does not seem to translate anything:

echo "after the" | moses -f moses.ini

BEST TRANSLATION: after|UNK|UNK|UNK the|UNK|UNK|UNK [11] ...

echo "after|after|IN the|the?P?????????????????((? MP?QI9M1Q%=8?????U9-?U9-?U9,????U9-?U9-?U9,?l??t????()Q??????????????????????????????????????????????)???????$???????????15???????????????????14??????????()Q?????????????????????????????????????????????????????)???????????????????$?????????15?????????????????????????)??????????????????????????????$??????????????)????????????????????????????????????????????????????)????????????(()Q??????????????????????



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

Message: 4
Date: Tue, 17 Mar 2015 09:48:16 -0400
From: Jeremy Gwinnup <jeremy@gwinnup.org>
Subject: [Moses-support] Support for Backoff Rule Tables?
To: moses-support@mit.edu
Message-ID: <38D4C9B8-2189-4235-BE8A-4BEEA987C1AD@gwinnup.org>
Content-Type: text/plain; charset=us-ascii

Hi,

Does moses currently support backoff rule tables in a similar fashion to backoff phrase tables? If so, do I need to specify a glue grammar for each rule table or is one sufficient?

Thanks!
-Jeremy


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

Message: 5
Date: Tue, 17 Mar 2015 19:05:14 +0500
From: Muhammad Danial Raza <muhammad.danial.raza@gmail.com>
Subject: Re: [Moses-support] Missing processPhraseTable in Release 3.0
To: Ertu?rul YILMAZ (B?LGEM) <yilmaz.ertugrul@tubitak.gov.tr>
Cc: moses-support <moses-support@mit.edu>
Message-ID:
<CAEQzcOf0vHo3gR726YmtkKBybpgG4d5WsookTAcay6=zP68Hmw@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Download the cmph library here <http://sourceforge.net/projects/cmph/> and
then build it using the instructions provided in the INSTALL file inside
the downloaded package.
where "/path/to/cmph" refers to the path where cmph library is installed

Then build the moses with the following command:
/usr/bin/bjam --with-boost=/home/ertugrul/tools/boost_1_55_0
--with-cmph=/path/to/cmph -j8
where "/path/to/cmph" refers to the path where cmph library is installed

Regards,
Danial

On Tue, Mar 17, 2015 at 6:32 PM, Ertu?rul YILMAZ (B?LGEM) <
yilmaz.ertugrul@tubitak.gov.tr> wrote:

> Hi All,
>
> I am upgrading our systems to the newest moses release (3.0), downloaded
> and compiled the sources as shown below
>
> git clone https://github.com/moses-smt/mosesdecoder.git --branch
> RELEASE-3.0 --single-branch RELEASE-3.0
> /usr/bin/bjam --with-boost=/home/ertugrul/tools/boost_1_55_0 -j8
>
> All completed with success. But I am missing the /bin/processPhraseTable
> executable in my executables. Has that executable been replaced by
> something else, or have I missed something along the way?
>
> Can someone provide information on this, pleaase?
>
> Thanks much
>
>
> *----------------------------------------------------------*
> *Ertu?rul YILMAZ*
>
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150317/d5293c9a/attachment.htm

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

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


End of Moses-support Digest, Vol 101, Issue 50
**********************************************

Tweet