It comes from the fact that you (1) don't have a solid background on Machine Learning/Statisitics, (2) your test set is completely/nearly our of domain from the training and dev set. On Wed, Jan 16, 2013 at 8:11 PM, chipsterchips <[email protected]>wrote:
> Hello everybody! > > I'm deploying the system on my machine with 4Gb Ram. Training process took > about 2 hours and the tuning process took about 12 hours. Then I tried to > translate text with tuned configuration file and with the original one to > compare their results. After some tests I didn't find any differences in > the quality of the translation. Can anybody explain me, what's the point of > tuining? > > Kirill > > _______________________________________________ > Moses-support mailing list > [email protected] > http://mailman.mit.edu/mailman/listinfo/moses-support > > -- * Best Regards, C. Hoang {Mimosa, SMT}@Addict *
_______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
