Hi,

I can't quite see how this would be beneficial, since the fullstop
should help with reordering and language model, but you can try
it out.

-phi

On Fri, Apr 22, 2011 at 8:19 AM, seagull <[email protected]> wrote:
> Hi there,
>
> I am using Moses to train English-Burmese corpus.
> In Burmese language, there is Sentence Marker as in English (Fullstop).
> Usually we train bi-text including them.
> Now i am thinking that I will remove those sentence markers in both src and
> trg corpus.
>
> I would like to know that:
> Is there any effect if sentence markers are excluded in training process?
> Might it decrease performance reordering model or language model?
>
> Kind Regards,
> Kyaw Kyaw Zin
>
> _______________________________________________
> 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