Thanks for your reply. This trick suggests that each word in the input sentence must repeat itself as a pseudo-factor, right?
If I get it right, an example input sentence could be: "this|this is|is a|a sentence|sentence for|for translation|translation .|." Also, I guess I need a realcase language model that should be defined as the surface form LM in moses.ini. In that case, I'm afraid that tuning for the LM weights is mandatory even for very domain-specific models... Panos On Saturday 08 of October 2011 15:55:44 Philipp Koehn wrote: > Hi, > > yes, you should prepare the output data with two factors, the lowercased > form and the recased form. You can then train a factored model with a > translation step (lowercase to lowercase) and a generation step (lowercase > to realcase). > > -phi > > On Fri, Oct 7, 2011 at 12:17 AM, Panos Kanavos <[email protected]> wrote: > > Hi all, > > > > In the paper "Factored translation models" (Philipp Koehn & Hieu Hoang, > > 2007), there is a section about integrated recasing with an extra > > generation step. Can this be done by adding an extra factor that > > identifies the case of each word on the target language? > > > > Thanks. > > > > Panos > > > > > > _______________________________________________ > > 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
