Hi Per,
I think that the explanation in this website:
http://rali.iro.umontreal.ca/rali/?q=en/node/1325 is quite useful. It helps
a lot to understand the structure and the content of each file generated by
OmegaT.
About the script, in the last release of bitextor we included a script
called "bitextor-builddics" (you can find the template of this script here:
https://svn.code.sf.net/p/bitextor/code/trunk/bitextor-builddics.in) which
uses GIZA++ to obtain a plain text bilingual dictionary, but only including
pairs of words fulfilling: a) both words occur at least 10 times in the
corpus, and b) the harmonic mean of their probabilities in both
probabilistic dictionaries (S -> T and T -> S) is higher than 0.2. If you
want to use this, I recommend you to use the version in the trunk, which
fixes some minor bugs still present in the release.
Best,
Miquel.
2014-02-17 14:21 GMT+01:00 Per Tunedal <[email protected]>:
> Hi Miquel,
> thank you for your informative answer. In deed I needed to create a
> coocurrence file.
> I did successfully create such a file with snt2cooc.out
>
> And GIZA++ has run successfully and made a lot of files in my home
> directory (!).
>
> How do I redirect the output to a more suitable folder? -outputpath ?
>
> Where can I find an explanation of the content of the files?
>
> I suppose the dictionary is in the translation table *.t3.final
> Any convenient way to extract plain text dictionaries (without going one
> step further and use Moses)?
> Some script available to decode the translation table by the using the
> vocabulary files *.vcb ?
>
> Yours,
> Per Tunedal
>
>
>
> On Mon, Feb 17, 2014, at 11:08, Miquel Esplà wrote:
>
> Hi Per,
>
> if I am not wrong, depending on how you compile GIZA++, it can generate
> the coocurrence files on-the-fly during alignment, or you may need to do so
> before running the alignment. Actually, I think that, with the standard
> compilation, you are in the second case. Have a look here:
> https://code.google.com/p/giza-pp/issues/detail?id=9 I hope the link will
> be helpful!
>
> Cheers,
>
> Miquel.
>
> 2014-02-17 10:30 GMT+01:00 Per Tunedal <[email protected]>:
>
>
> Hi,
> I tried the procedure described at
> http://wiki.apertium.org/wiki/Using_GIZA%2B%2B to get a rough
> dictionary, but encountered the following error in the last step:
>
> ERROR: NO COOCURRENCE FILE GIVEN!
>
> Is one step missing in the procedure?
>
> Yours,
> Per Tunedal
>
>
>
> ------------------------------------------------------------------------------
> Android apps run on BlackBerry 10
> Introducing the new BlackBerry 10.2.1 Runtime for Android apps.
> Now with support for Jelly Bean, Bluetooth, Mapview and more.
> Get your Android app in front of a whole new audience. Start now.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=124407151&iu=/4140/ostg.clktrk
> _______________________________________________
> Apertium-stuff mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/apertium-stuff
>
>
> ------------------------------------------------------------------------------
> Android apps run on BlackBerry 10
> Introducing the new BlackBerry 10.2.1 Runtime for Android apps.
> Now with support for Jelly Bean, Bluetooth, Mapview and more.
> Get your Android app in front of a whole new audience. Start now.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=124407151&iu=/4140/ostg.clktrk
> *_______________________________________________*
> Apertium-stuff mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/apertium-stuff
>
>
>
> ------------------------------------------------------------------------------
> Android apps run on BlackBerry 10
> Introducing the new BlackBerry 10.2.1 Runtime for Android apps.
> Now with support for Jelly Bean, Bluetooth, Mapview and more.
> Get your Android app in front of a whole new audience. Start now.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=124407151&iu=/4140/ostg.clktrk
> _______________________________________________
> Apertium-stuff mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/apertium-stuff
>
>
------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
