Moses-support Digest, Vol 99, Issue 51

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: Starting words are left untranslated (Hieu Hoang)
2. about the scores in run*.best100.out (Arefeh Kazemi)
3. Missing processPhraseTable binary from bin directory (Mario Diana)
4. Re: Missing processPhraseTable binary from bin directory
(Hieu Hoang)


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

Message: 1
Date: Thu, 22 Jan 2015 17:24:45 +0000
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] Starting words are left untranslated
To: CHATZITHEODOROU Konstantinos <dinosaxon@gmail.com>,
"moses-support@mit.edu" <moses-support@mit.edu>
Message-ID: <54C1325D.2020305@gmail.com>
Content-Type: text/plain; charset="windows-1252"

is it 'cos it's uppercased?

are you sure your input test set doesn't have the Windows return
character (0x13)? Get rid of it by doing dos2unix

On 22/01/15 17:11, CHATZITHEODOROU Konstantinos wrote:
>
> Hello Moses,
>
> I have trained an EN-IT phrase based MT Engine with a very high BLEU
> score. I have also checked the test output and the MT quality is
> pretty good.
>
> Now, I am trying to translate another document, different from the
> test set but I realised that a large number of words are left
> untranslated even if in the phrase translation table these words
> appear having their own entries (phrase length =1). Moreover, I
> observed that most of the times, the first word of each sentence is
> untranslated. I run moses decoder with -output-unknowns option.
>
> Here are some examples:
>
>
> Think and respond to the questions . - > Think e
> rispondere alle domande .
>
> I think . -> Io penso .
>
>
> One more issue. The word "real" is translated as "rear" (posteriore)
> in Italian. Again, I checked the phrase translation table entries and
> there is no matching between these two words. The only match is in
> the reordering table.
>
>
> Do you have any idea what is going wrong and what I can do?
>
>
> Thank you for your time.
>
>
> Best,
>
> Konstantinos
>
>
>
> _______________________________________________
> 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/20150122/5d4b17e0/attachment-0001.htm

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

Message: 2
Date: Thu, 22 Jan 2015 21:00:22 +0000 (UTC)
From: Arefeh Kazemi <arefeh_kazemi@yahoo.com>
Subject: [Moses-support] about the scores in run*.best100.out
To: Moses-support <moses-support@mit.edu>
Message-ID:
<1718111479.3164319.1421960422061.JavaMail.yahoo@jws10086.mail.ne1.yahoo.com>

Content-Type: text/plain; charset="utf-8"

Hi
I've implemented a feature with 4 scores in moses-chart. when I debug the code on my system, every thing is OK and the scores are calculated right, but when I run mert to tune the weights (on a cluster), I get wrong scores for my feature in run*.best.100 files. for example my feature could have the maximum value of 60 but for some instances I get values more than 300.
does anyone know why this happens?
Regards
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150122/00ef20d4/attachment-0001.htm

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

Message: 3
Date: Thu, 22 Jan 2015 21:55:30 +0000
From: Mario Diana <mdiana@technicallycreative.com>
Subject: [Moses-support] Missing processPhraseTable binary from bin
directory
To: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<SN2PR07MB061D43D7E1477D8047A3B74C6490@SN2PR07MB061.namprd07.prod.outlook.com>

Content-Type: text/plain; charset="us-ascii"

Hello,

I have been running through the tutorial here:

http://www.statmt.org/moses/?n=moses.baseline

I have gotten as far as the "Tuning" section and have completed that successfully. I am able to run Moses from the command line and translate French sentences into English. I am having a problem with the next section, where we "binarise" the phrase-table.

The command in the tutorial reads:

~/mosesdecoder/bin/processPhraseTable \
-ttable 0 0 train/model/phrase-table.gz \
-nscores 5 -out binarised-model/phrase-table

However, I am missing the processPhraseTable binary. That seems strange to me.

If I run ls on my directory, I get the following:

[mdiana@localhost ~]$ ls ~/mosesdecoder/bin/pr*
/home/mdiana/mosesdecoder/bin/pro
/home/mdiana/mosesdecoder/bin/processLexicalTable
/home/mdiana/mosesdecoder/bin/prunePhraseTable

Can it be the case that processPhraseTable is deprecated and that the tutorial hasn't caught up? Is "pro" the name of the binary now (abbreviated for some reason)? Did I make some common mistake in setting up my environment? I just don't get it and would appreciate any hints. I would like to binarise my phrase table.

Thank you,

Mario

::::::::::::::::::::::::::::::::::::::::::::::::::
Mario Diana
Software Developer
Technically Creative Inc.
Simplifying IT Solutions
Office: 845.725.7883
mdiana@technicallycreative.com
www.TechnicallyCreative.com
::::::::::::::::::::::::::::::::::::::::::::::::::




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

Message: 4
Date: Thu, 22 Jan 2015 21:58:26 +0000
From: Hieu Hoang <Hieu.Hoang@ed.ac.uk>
Subject: Re: [Moses-support] Missing processPhraseTable binary from
bin directory
To: Mario Diana <mdiana@technicallycreative.com>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAEKMkbj0OiAROQ0=iaY1LQZ_w7d=AUGAxyNT=L=3cmYTjgqhgQ@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

fyi
https://www.mail-archive.com/moses-support@mit.edu/msg11367.html


On 22 January 2015 at 21:55, Mario Diana <mdiana@technicallycreative.com>
wrote:

> Hello,
>
> I have been running through the tutorial here:
>
> http://www.statmt.org/moses/?n=moses.baseline
>
> I have gotten as far as the "Tuning" section and have completed that
> successfully. I am able to run Moses from the command line and translate
> French sentences into English. I am having a problem with the next section,
> where we "binarise" the phrase-table.
>
> The command in the tutorial reads:
>
> ~/mosesdecoder/bin/processPhraseTable \
> -ttable 0 0 train/model/phrase-table.gz \
> -nscores 5 -out binarised-model/phrase-table
>
> However, I am missing the processPhraseTable binary. That seems strange to
> me.
>
> If I run ls on my directory, I get the following:
>
> [mdiana@localhost ~]$ ls ~/mosesdecoder/bin/pr*
> /home/mdiana/mosesdecoder/bin/pro
> /home/mdiana/mosesdecoder/bin/processLexicalTable
> /home/mdiana/mosesdecoder/bin/prunePhraseTable
>
> Can it be the case that processPhraseTable is deprecated and that the
> tutorial hasn't caught up? Is "pro" the name of the binary now (abbreviated
> for some reason)? Did I make some common mistake in setting up my
> environment? I just don't get it and would appreciate any hints. I would
> like to binarise my phrase table.
>
> Thank you,
>
> Mario
>
> ::::::::::::::::::::::::::::::::::::::::::::::::::
> Mario Diana
> Software Developer
> Technically Creative Inc.
> Simplifying IT Solutions
> Office: 845.725.7883
> mdiana@technicallycreative.com
> www.TechnicallyCreative.com
> ::::::::::::::::::::::::::::::::::::::::::::::::::
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>



--
Hieu Hoang
Research Associate
University of Edinburgh
http://www.hoang.co.uk/hieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150122/509744de/attachment.htm

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

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


End of Moses-support Digest, Vol 99, Issue 51
*********************************************

Tweet