Dear Experts, I am working on a Spanish to Catalan and Catalan to Spanish decoder. I've done the whole training for the Spanish to Catalan version and it works fine, but when I try the inverse process it fails.
As I have seen, the perl script 'train-factored-phrase-model' creates the folders giza.es-ca and giza.ca-es on its GIZA++ step. I was wondering if I should make a training for the Spanish to Catalan decoder, and, if I should, if there is a step I can skip on the training for the Spanish to Catalan decoder. Best Regards, Miguel _______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
