6 2.718 ||| 0-0 1-1 1-2 2-3 ||| 1 2 1
>
>
> What do the values 1 2 1 are for ?
>
> Also does moses use the alignments 0-0 and 1-1 etc to give translations.
>
> Thanks
> Megha
> -- next part --
> An HTML attachment was scrubbed...
> URL:
> ht
I made an implementation of a transliteration package from English-2-Indian
Languages(including Hindi). I guess there are other packages also available.
link to package : https://github.com/ksingla025/Motech_Transliterate/
you can use it as a post-processing tool, for words which are OOV or
un-tra
Hello,
I have a question concern Mert-moses: Is it possible in Moses to do Mert on
XML files?
Is there any option to do this, or should i delete the XML tags manually
before doing MERT?
Thank you in advance
Bests
Cyrine
___
Moses-support mailing list
Hi,
Is there a module to transliterate OOV words in Moses? Or any standalone
transliteration toolkit to do this?
I am trying to translate some English data into Hindi. Our system is based
on the data provided for WMT this year and some additional in-domain data.
But the number of OOVs is still ma
the 1 2 1 are the counts. They are used for debug and can be ignored, they
aren't used for decoder.
the word alignments 0-0 and 1-1 are used by some new feature functions but
not by the basic feature functions. If you don't want to create them, then
leave them out
On 8 June 2014 12:37, Megha Jhu
Hello,
I am trying to do phrase-table triangulation to generate a new
phrase-table. But I am not being able to figure what the last 3 values in
the phrase-table signify. પીઠના
for ex :
દુખાવાથી પીડિત ||| पाठदुखीचे शिकार ||| 0.00864127194755 8.337659958e-08
0.00432064057394 1.13306222125e-06 2.718
