Francis Tyers <[email protected]> writes: > Hello all, > > What do you think about having a new mode (yes, *groan*, a new mode) > for lt-proc where we can generate keeping the lexical form, e.g. > > Input: > > ^cantar<vblex><pres><p1><sg>$ > ^de<pr>$ > > Normal output '-g': > > canto > ~de > > Output with '-k -g' mode: > > ^canto/cantar<vblex><pres><p1><sg>$ > ^~de/de<pr>$ > > This would be teamed with a '-k -p' mode for postgeneration which would > strip the analysis: > > canto > ~de > > Why would this be useful? Well, I could imagine you could do stuff with > the analysis. Like with a language model or something. Hmm, if, e.g. you > wanted to have a module which posteditted prepositions or something and > you wanted to train it on lexical forms as well as surface forms. > > Fran
+1. Could be useful in testvoc as well. Also, lt-proc -a already goes from <l> to both <l> and <r> – going from <r> to both <l> and <r> would make -g more symmetric :) -Kevin ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Apertium-stuff mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/apertium-stuff
