Dear Per, Seems that you have lots of questions that you've put question mark even after your name. ;-) The output that you got seems fine. All you need is to run the fast aligner in two directions (forward and reverse; without and with -r option) and get the symmetrized alignment using the perl script that you can find it here: mosesdecoder / scripts / ems / support / symmetrize-fast-align.perl
So, basically you just need to store the outputs of fast_align and symmetrize them. Moses just needs this symmetrize word alignment to extract the phrase table and other stuff. I am pretty sure that in EMS there is the option for using fast_align instead of (m)giza++, but since I don't use EMS, I don't know how you can do it. You can take a look at the following page: http://www.statmt.org/moses/?n=FactoredTraining.EMS#ntoc13 It might help you in figuring out how to use it. Enjoy your experiments, Amin On 02/17/2014 07:45 PM, Per Tunedal wrote: Hi, I've run Fast Align and got the output: 0-0 2-1 0-0 2-1 3-2 5-3 0-0 1-1 2-2 2-3 0-0 1-1 3-2 2-3 4-4 5-5 7-6 7-7 0-0 1-1 2-2 0-0 2-1 0-0 2-1 4-2 5-3 0-0 1-1 2-2 2-3 0-0 1-1 3-2 2-3 4-4 5-5 7-6 7-7 0-0 1-1 2-2 Fine, but there must be some other files somewhere, mustn't it? What word is encoded as 4 etc? Where are the files saved? What are they? How does Moses find them? Yours, Per Tunedal? _______________________________________________ Moses-support mailing [email protected]http://mailman.mit.edu/mailman/listinfo/moses-support
_______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
