seems fine to me. Using your data: ./scripts/recaser/train-truecaser.perl --model model --corpus ~/Downloads/english.tok.en i get: # head model 's (1/1) the (12/13) THE (1) that (1/1) , (4/4) on (1/1) earth (1/1) Let (1/1) day (1/1) darkness (1/1) earth. (2/2)
On 7 September 2013 19:04, Deeptiman Pattnaik <[email protected]>wrote: > Hello, > > I am very new Moses, so i was applying the some of the commands from moses > mannual but got following error. > > I have started doing the Corpus preparation in Moses. But getting some > error. > > you can see in the Moses manual Page No-: 27 , Corpus Preparation > > Moses Manual -: http://www.statmt.org/moses/manual/manual.pdf > > > *Tokenisation* > > i have successfully completed Tokenisation of English Corpora, i got > two output file > > english.tok.en > > > *Truecaser > * > > When i proceed for the truecase of the English corpora , we typed > following commands. > > > ~/mosesdecoder/scripts/recaser/train-truecaser.perl \ > > --model ~/mosesdecoder/corpus/truecase-model.en --corpus \ > > ~/mosesdecoder/corpus/english.tok.en > > but we got the error as follows-: > > train-truecaser.perl --model truecaser --corpus cased > [--possiblyUseFirstToken] at > /home/deeptiman/mosesdecoder/scripts/recaser/train-truecaser.perl line 16. > > > i have searched in net , i have found that the first character every > sentence should be Capitalized and our corpora also Capitalized but we > can't able to execute this command. > > > Please go through this problem and help us out. > > _______________________________________________ > Moses-support mailing list > [email protected] > http://mailman.mit.edu/mailman/listinfo/moses-support > > -- Hieu Hoang Research Associate University of Edinburgh http://www.hoang.co.uk/hieu
_______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
