> I'm trying to replicate an experiment done with an older version of > Moses, against the latest version. Everything goes identically up to > tuning, where the newer version starts with a lower BLEU and runs > only 8 as opposed to 19 iterations of MERT, resulting in a lower > evaluation score.
Hi John -- I guess Hieu is looking at the deeper problem here, but to address the above comment -- the number of iterations is usually not correlated with performance, so I wouldn't worry too much about this. Cheers Adam -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. _______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
