Hi, each possible translation of the input sentence is scored by a few different component models, such as the language model, the direct phrase translation probabiity, etc. Tuning sets weights for these component models based on how important they are to achieve good translations.
Tuning is described here: http://www.statmt.org/moses/?n=FactoredTraining.Tuning The MERT paper by Franz Och http://aclweb.org/anthology//P/P03/P03-1021.pdf introduced these ideas into statistical machine translation. -phi On Tue, Jan 14, 2014 at 5:26 AM, nadeem khan <[email protected]> wrote: > Hi all; > > I want to know some details about the Moses weights that are generated > after training step under moses.ini config file and also we use tuning step > to improve those weights. > Where & how actually these weights are being used? and how decoder take > those weights while translating a input sentence? Can someone elaborate the > use of these weights with example. > What if we modified those weights manually without using tuning? > > Regards > nadeem khan > > _______________________________________________ > 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
