Hi, the script, as you invoke it, expects two files: /home/tjr/corpus/multiUN.ar-en.true.ar /home/tjr/corpus/multiUN.ar-en.true.en
Do they exist? How big are they? -phi On Mon, Jul 15, 2013 at 1:53 PM, Heidi Heweidy <[email protected]> wrote: > I have just finished truecasing arabic-english corpora fromthe multiUN > parallel text > corpora and now I was at the cleaning step but I had an error when typing > this: > ~/mosesdecoder/scripts/training/clean-corpus-n.perl > ~/corpus/multiUN.ar-en.true ar en ~/corpus/multiUN.ar-en.clean 1 80 > > The result was: > clean-corpus.perl: processing /home/tjr/corpus/multiUN.ar-en.true.ar & .en > to /home/tjr/corpus/multiUN.ar-en.clean, cutoff 1-80 > Input sentences: 0 Output sentences: 0 > > Why would this be? Can it be because of the 'ar' symbol as it is undefined > in the Moses system or sth else? > > _______________________________________________ > 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
