Miles Osborne wrote: > i'd check to see how unknown words are handled in either the SRILM > or in IRSTLM --that may explain the differences
Ah, good suggestion, thanks - OOV is very high in this data. > (as for the size of a tuning set, the more the better; right now > i'm doing Europarl runs (Fr -- En) and our dev set has 2k sentences.) Hmm, okay. I thought 200 or so was common. Sadly, we have a very limited amount of in-domain data, and there's the usual tradeoff between having enough in the tuning set for MERT to do a good job, and having enough devtest to get good significance between different experiments. We might have to do cross-validation of some kind ... - JB _______________________________________________ Moses-support mailing list Moses-support@mit.edu http://mailman.mit.edu/mailman/listinfo/moses-support