Hi Laxmi,

have a look at the moses website. Your corpus should be of the format 
explained here: 
http://www.statmt.org/moses/?n=FactoredTraining.PrepareTraining

You can find the training parameters at: 
http://www.statmt.org/moses/?n=FactoredTraining.TrainingParameters

and some specific examples, e.g.: 
http://www.statmt.org/moses/?n=Moses.FactoredTutorial#ntoc1
The "train-factored-phrase-model.perl" script trains a new system for 
you. (So there is no adding of a corpus but training your very own 
system with your corpus.) You need to specify at least where the corpus 
is located (--corpus path), where your mt-system should be stored 
(--root-dir path), and which file extension you use for the corpora 
files (--f x --e y). So corpus files e.g. "corpus.nepali" and 
"corpus.english" should be specified as --f nepali --e english for 
translation from nepali to english. Furthermore the location of you 
language model (--lm) needs to be specified.
You need to train the LM separately. Again, look at the website for 
further information: 
http://www.statmt.org/moses/?n=FactoredTraining.BuildingLanguageModel

Best is to follow the steps on http://www.statmt.org/moses_steps.html 
(Part II) to train the whole system.

Good luck.
Christine


Laxmi Khatiwada schrieb:
> Hello
>
> I have already installed moses and support tools.
> I can run en fr sample model.
>
> Now I want to add new corpus in moses. Nepali corpus. But I have no 
> Idea to add new language pair.
> Can you suggest me the right process?
>
> Laxmi
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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

Reply via email to