Hi Joerg. 

I can confirm that in my case it was a problem with the phrase table. I checked 
it again today and I found some phrases that had malformed tokens in the target 
side. They were like "wrd1| wrd2|", with the second factor missing. I made 
those phrase tables with different scripts so it was my fault. 

Anyway, the error message give the "right clue", it was indeed a problem with 
malformed factors, but it could be either in the input file or in the phrase 
table (source side or target side).

Thank you all. 

Bye.


----- Mensaje original -----
De:Joerg Tiedemann <[email protected]>
Para:Carlos Henriquez <[email protected]>
CC:"[email protected]" <[email protected]>
Enviado:lunes, 22 de noviembre, 2010 7:04:28
Asunto:Re: [Moses-support] Malformed input at (blank)

Several of my students had te same problem recently when playing
around with factored models. I also think that the error message must
be misleading. After filtering the phrase table according to the input
everything worked fine. It must have something to do with the phrase
table ...

Jörg


On Mon, Nov 22, 2010 at 5:52 PM, Carlos Henriquez
<[email protected]> wrote:
> Hi all.
>
> Did you solve this problem?
>
> I'm having the same issue right now and I have not found an answers for it.
>
> [ERROR] Malformed input at
>  Expected input to have words composed of 2 factor(s) (form FAC1|FAC2|...)
>  but instead received input with 1 factor(s).
>
>> This error message shows up at just the beginning
>> of the decoding phase after loading phrase table before decoding the first
>> input.
>
> Could the problem be in the translation table? regardless of the "malformed 
> input" message. I said this because I changed the translation-table in the 
> moses.ini file for an old one I had and the decoder started just fine with 
> the same input file.
>
> I found some "[" and "]" in my phrases and replaced them but it did not solve 
> the problem. Are there any other reserved character? I made this translation 
> table from scratch, combining phrases and scores from two "moses made" 
> ttables.
>
> Are you sure the problem is in the input file? Why does moses work with the 
> same input but a different ttable?
>
> Btw, both ttable are 0-0,1 (word|tag). I've checked and no pipes appear in 
> the input nor in the source side of the translation table. The old one I 
> mentioned before was "moses made".
>
> --
>
> De: Barry Haddow <[email protected]>
> Para: [email protected]
> CC: [email protected]
> Enviado: martes, 26 de octubre, 2010 8:02:28
> Asunto: Re: [Moses-support] Malformed input at (blank)
>
> Hi Mehmet
>
> Have you got some unprintable characters at the start of your input file? Try
> running xxd to see what's there.
>
> regards
> Barry
>
> On Tuesday 26 October 2010 18:54, Mehmet Tatlıcıoğlu wrote:
>> No luck with either way. This error message shows up at just the beginning
>> of the decoding phase after loading phrase table before decoding the first
>> input.
>>
>> --
>> Mehmet Tatlıcıoğlu
>> Cell Phone: +90 532 201 85 64
>>
>> On 26 October 2010 17:57, Ondrej Bojar <[email protected]> wrote:
>> > Hi,
>> >
>> > one more little issue could be caused by several spaces in a row or
>> > spaces at the beginnings and ends of lines.
>> >
>> > O.
>> >
>> > Hieu Hoang wrote:
>> > > that's strange. Check which input line its failing on.
>> > >
>> > > Also, check it doesn't have [ or ] characters as these are now also
>> > > reserved for doing hierarchical/syntax models.
>> > >
>> > > On 26 October 2010 15:32, Mehmet Tatlıcıoğlu <[email protected]
>> > > <mailto:[email protected]>> wrote:
>> > >
>> > >     Hi,
>> > >     As far as I remember, there was some people around telling that the
>> > >     decoder, sometimes, dies with the following error message,
>> > >     [ERROR] Malformed input at
>> > >       Expected input to have words composed of 1 factor(s) (form
>> > >     FAC1|FAC2|...)
>> > >       but instead received input with 0 factor(s).
>> > >     Aborted
>> > >     Exit code: 134
>> > >     The decoder died. CONFIG WAS -w 0.000000 -lm 0.111111 -d 0.111111
>> > >     0.111111 0.111111 0.111111 0.111111 0.111111 0.111111 -tm 0.033333
>> > >     0.022222 0.033333 0.022222 0.000000
>> > >
>> > >     Moreover, according to the pages I searched in, the possible
>> > >     solution recommended was to clean the input data from '|'
>> > >     characters. However, some people having this error said they didn't
>> > >     have any '|' character indeed. Now, I had the same error message,
>> > >     and I don't have neither '|' character nor any consecutive blank
>> > >     characters (nor any white space at the beginning of a line) in my
>> > >     input data.
>> > >
>> > >     So, does this error message show up due to any other reason?
>> > >     Possibly, a bug?
>> > >
>> > >     I can run a decoder for a translation system, a -> b successfully.
>> > >     However, trying to do it as b -> a with the same data, I get the
>> > >     error message above.
>> > >
>> > >     Thanks in advance...
>> > >
>> > >     --
>> > >     Mehmet Tatlıcıoğlu
>> > >     Cell Phone: +90 532 201 85 64
>> > >
>> > >     _______________________________________________
>> > >     Moses-support mailing list
>> > >    [email protected] <mailto:[email protected]>
>> > >    http://mailman.mit.edu/mailman/listinfo/moses-support
>> > >
>> > >
>> > >
>> > > -----------------------------------------------------------------------
>> > >-
>> > >
>> > > _______________________________________________
>> > > Moses-support mailing list
>> > > [email protected]
>> > > http://mailman.mit.edu/mailman/listinfo/moses-support
>
> --
> 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
>
>
>
>
> _______________________________________________
> Moses-support mailing list
> [email protected]
> http://mailman.mit.edu/mailman/listinfo/moses-support
>



-- 
**********************************************************************************
Jörg Tiedemann                                 http://stp.lingfil.uu.se/~joerg/



      

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

Reply via email to