Hi Horia

The new phrase table format looks like this:

! " , mientras ||| ! " while ||| 1 0.00128018 1 0.187558 2.718 ||| 0-0 1-1 2-2 
3-2 ||| 1 1

ie, it's
source ||| target ||| scores ||| word alignment ||| count

As far as I recall, only the first three are mandatory,

cheers - Barry

On Tuesday 20 December 2011 12:57:10 Horia Cucu wrote:
> Hi Barry,
> 
> I'm trying to figure out how to convert the old phrase table into the new
> format and I have a short question. Here's how my current phrase table
> looks like:
> ...
> - a c o l o ||| a k o l o ||| () (0) (1) (2) (3) (4) ||| (1) (2) (3) (4)
> (5) ||| 0.0117647 0.0101026 1 0.621085 2.718
> - a c o l ||| a k o l ||| () (0) (1) (2) (3) ||| (1) (2) (3) (4) |||
> 0.00378788 0.0101027 1 0.654066 2.718
> ...
> 
> I've read about the new format and I think I should transform this lines
> into:
> ...
> - a c o l o ||| a k o l o ||| 1-0 2-1 3-2 4-3 5-4 ||| 0-1 1-2 2-3 3-4 4-5
> 
> ||| 0.0117647 0.0101026 1 0.621085 2.718
> 
> - a c o l ||| a k o l ||| 1-0 2-1 3-2 4-3 ||| 0-1 1-2 2-3 3-4 |||
> 0.00378788 0.0101027 1 0.654066 2.718
> ...
> 
> Is this right or am I getting something wrong?
> Thanks,
> Horia
> 
> On 16 December 2011 11:39, Barry Haddow <[email protected]> wrote:
> > Hi Horia
> >
> > The phrase table format has changed so you might have to recreate or
> > convert
> > it. You should first try changing the phrase-table line in your moses.ini
> > from
> >
> > 0 0 5 <ttable-path>
> >
> > to
> >
> > 0 0 0 5 <ttable-path>
> >
> > The leading '0' signifies that the table is in text format. The new text
> > format
> > for phrase tables is given here
> > http://www.statmt.org/moses/?n=FactoredTraining.ScorePhrases
> > so you should be able to convert your old one with a simple script,
> >
> > cheers - Barry
> >
> > On Thursday 15 December 2011 14:05:09 Horia Cucu wrote:
> > > Hi,
> > >
> > > I have a phrase table created with a version of moses compiled in 2007.
> > > I've just installed moses on my computer and tried to decode using this
> >
> > old
> >
> > > phrase table and I'm stuck here:
> > >
> > > ...
> > > ...
> > > Loading lexical distortion models...have 0 models
> > > Start loading LanguageModel
> > > /home/cucu/speechRoot/tools/wordsPhonetization/dist/dict.train.ph.lm :
> > > [0.000] seconds
> > > Finished loading LanguageModels : [0.000] seconds
> > > Warning: Phrase table specification in old 4-field format. Assuming
> >
> > binary
> >
> > > phrase tables (type 1)!
> > > Start loading PhraseTable
> >
> > /home/cucu/speechRoot/tools/wordsPhonetization/filteredTable/phrase-table
> >.0
> >
> > > -0.1
> > >
> > > : [0.000] seconds
> > >
> > > filePath:
> >
> > /home/cucu/speechRoot/tools/wordsPhonetization/filteredTable/phrase-table
> >.0
> >
> > > -0.1 Finished loading phrase tables : [0.000] seconds
> > > IO from STDOUT/STDIN
> > > Created input-output object : [0.000] seconds
> > > Translating line 0  in thread id 3074714480
> > > Translating: c a s ă
> > >
> > > bin ttable does not exist -> create it
> > > distinct source phrases: 19 distinct first words of source phrases: 8
> > > number of phrase pairs (line count): 41
> > > reading bin ttable
> > > size of OFF_T 8
> > > binary phrasefile loaded, default OFF_T: -1
> > > Check weightT.size() == scoreVector.size() failed in
> > > moses/src/TargetPhrase.cpp:155
> > > ./translateWordList.sh: line 16: 21578 Aborted                 moses -f
> > > filteredTable/moses.ini < $WORDLIST.gr > $WORDLIST.ph
> > >
> > > any ideas?
> > > horia
> >
> > --
> > Barry Haddow
> > University of Edinburgh
> > +44 (0) 131 651 3173
> >
> > --
> > The University of Edinburgh is a charitable body, registered in
> > Scotland, with registration number SC005336.
> 
 
--
Barry Haddow
University of Edinburgh
+44 (0) 131 651 3173

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


_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to