Hi. The nan errors do not necessarily mean that the moses.ini will be corrupt, i.e. the moses.ini file may seem correct, but the translation model built may be of poor(er than expected) quality. Usually, nan problems during training mean that the training procedure did not run completely well. It might be because of a non-filtered training set (observe, in the moses howto, the step "filter out long sentences"), or it might also be because the alignments you are using are in some way distorted.
It would help if you could tell us where the nan errors appeared, but it is a good idea to find out why you get those errors, since it might be affecting the final translation quality. Or not, but just in case. Good luck, Germán Sanchis P.S.: just in case, I am referring to the howto in http://www.statmt.org/wmt08/baseline.html Quoting chaitu chaitu <[EMAIL PROTECTED]>: > Hello all > I just finished training and it created phrasae table and moss.ini > > I am not sure if my moses.ini is correct or is currupted as I got > several errors during the training. The error is some thing like nan > nan nan nan nan nan and so on. > > Here is my moses.ini file > Can any one say if it is correct or do I need to train again. > ######################### > ### MOSES CONFIG FILE ### > ######################### > > # input factors > [input-factors] > > > # mapping steps > [mapping] > 0 T 0 > > # translation tables: source-factors, target-factors, number of scores, file > [ttable-file] > 0 0 5 /home/moses/webapps/working-dir/model/phrase-table.0-0.gz > > # no generation models, no generation-file section > > # language models: type(srilm/irstlm), factors, order, file > [lmodel-file] > 0 0 5 working-dir/lm/europarl.lm > > > # limit on how many phrase translations e for each phrase f are loaded > # 0 = all elements loaded > [ttable-limit] > 20 > > # distortion (reordering) files > [distortion-file] > 0-0 msd-bidirectional-fe 6 > /home/moses/webapps/working-dir/model/reordering-table.msd-bidirectional-fe.0.5.0-0.gz > > # distortion (reordering) weight > [weight-d] > 0.3 > 0.3 > 0.3 > 0.3 > 0.3 > 0.3 > 0.3 > # language model weights > [weight-l] > 0.5000 > > > # translation model weights > [weight-t] > 0.2 > 0.2 > 0.2 > 0.2 > 0.2 > > # no generation models, no weight-generation section > > # word penalty > [weight-w] > -1 > > [distortion-limit] > 6 > > > > Thanking You in anticipation > > > > > --------------------------------- > Unlimited freedom, unlimited storage. Get it now ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. _______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
