Hi everybody,

this is not exactly a question about Moses, but I think this is the most
adequate list to get some help. I am, trying to train GIZA++ with a corpus
and, then, to use the models generated to align another corpus. I tried to
use the "testcorpusfile" option in GIZA++, but it seems that it fails when a
word in the test corpus does not appear in the train corpus (this is, it is
not in the .vcb file). I though to add these words to the .vcb file and
assign to them value 0 in the third column (the number of times that these
words are found in the training corpus). This seems to work, but I am not
sure if it is correct... have anybody ever tried to do this?

Thanks in advance. Kind regards,

Miquel.
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to