Hi John I hit against this problem in binarisation too. I think it's caused by the phrase count bug that I fixed last Thursday. I'm going to retry with the updated scripts,
cheers - Barry On Thursday 07 Jul 2011 18:41:04 John Morgan wrote: > Thanks Hieu, > The rule table binarization failed. Those files were empty (except > for some symbols). > I'm attaching the error file for the binarization. > jjm > > On 7/7/11, Hieu Hoang <[email protected]> wrote: > > that's weird. > > > > has the rule table has been binarized? There should be the following > > files in > > > > /home/john/TF435/d2e/baseline/hierarchical/evaluation/filtered.trials.8/p > >hrase-table.0-0.1.1.bin Vocab.dat > > TargetInd.dat > > TargetColl.dat > > Source.dat > > Misc.dat > > the file Misc.dat should look a bit like > > Version 3 > > NumSourceFactors 1 > > NumTargetFactors 1 > > NumScores 1 > > RootNodeOffset 300 > > the assert error indicate some of this info isn't there > > > > On 07/07/2011 20:27, John Morgan wrote: > >> Hello, > >> I'm trying to decode using a system constructed with the ems. I've > >> attached a file with > >> the commandline and trace. Is there a problem with the word penalty? > >> I'm using > >> > >> > >> _______________________________________________ > >> 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 > _______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
