Hi,

sure, no problem.

If you train a model with train-model.perl, just skip the
first two or three training steps (--first-step N) and point
to the GIZA++ (--giza-e2f and ---giza-f2e) or symmetrized
word alignment file (--alignment-file).

Similarly, if you use experiment.perl, just point to your
existing files.

-phi

2012/1/29 Onur Görgün <[email protected]>

>  Hi,
>
> I'm having problems with installing Moses and GIZA and SRILM together, but
> I already have language model and GIZA++ files on hand. Is it possible to
> train
> a phrase-based translation model with these files and without using
> supporting tools?
>
> Thank you,
>
> Onur
>
> _______________________________________________
> 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