Dear Moses group, you may already have received my earlier message on my attempts to install Moses on a Mac, which ended with the following error:
Executing: cat ./model/extract.0-0.o.part* > ./model/extract.0-0.o cat: ./model/extract.0-0.o.part*: No such file or directory Exit code: 1 Died at bin/moses-scripts/scripts-20080811-1801/training/train-factored-phrase-model.perl line 891. It seems that this problem was encountered before by Nguyen Tien Huy. See below Philipp Koehn's advice given at that time. I have looked at the contents of model directory and it's similar to the one below. Mine is: -rw-r--r-- 1 lliohump lliohump 121523720 Aug 12 14:01 aligned.0.en -rw-r--r-- 1 lliohump lliohump 132506103 Aug 12 14:00 aligned.0.it -rw-r--r-- 1 lliohump lliohump 0 Aug 12 13:58 aligned.grow-diag-final-and -rw-r--r-- 1 lliohump lliohump 0 Aug 12 14:06 extract.0-0.o -rw-r--r-- 1 lliohump lliohump 5044265 Aug 12 14:05 lex.0-0.f2n -rw-r--r-- 1 lliohump lliohump 5044265 Aug 12 14:05 lex.0-0.n2f The contents of giza.en-it is: -rw-r--r-- 1 lliohump lliohump 153243919 Aug 12 13:57 en-it.A3.final.gz -rw-r--r-- 1 lliohump lliohump 361711995 Aug 12 07:26 en-it.cooc -rw-r--r-- 1 lliohump lliohump 1906 Aug 12 07:26 en-it.gizacfg and the contents of giza.it-en is: -rw-r--r-- 1 lliohump lliohump 152598458 Aug 12 07:18 it-en.A3.final.gz -rw-r--r-- 1 lliohump lliohump 362039740 Aug 12 00:39 it-en.cooc -rw-r--r-- 1 lliohump lliohump 1906 Aug 12 00:39 it-en.gizacfg It would therefore seem appropriate for me to proceed with Philipp's advice: " If that is OK, check, if you can run the symal command by hand. (check for "Executing: /.../scripts-20070109-1725/training/symal/giza2bal.pl") in the STDERR of the training script. " The only problem is I don't know how to do this! I take it that stderr prints error messages to the terminal? Unfortunately, I can only save a certain amount of text output to the terminal, and I can't find anything related to symal output to the screen in that portion. In fact, what I do have are the following lines over and over again: Use of uninitialized value in scalar chomp at bin/moses-scripts/scripts-20080811-1801/training/train-factored-phrase-model.perl line 790. Use of uninitialized value in split at bin/moses-scripts/scripts-20080811-1801/training/train-factored-phrase-model.perl line 793. I don't know understand what this means. train-factored-phrase-model.perl line 790 says: my $a = <A>; chomp($a); which looks like it assigns variable a with value <A>, and I read somewhere that chomp 'simply removes any erroneous line breaks and spacing from the end of a string', so I can't see why this should cause an error. Maybe the error comes from the previous line? my @FOREIGN = split(/ /,$f); And yet train-factored-phrase-model.perl line 793 says: foreach (split(/ /,$a)) which wouldn't work if there's a problem with $a..... Does this information help shed some light on where I'm going wrong with this? Do I need to amend train-factored-phrase-model.perl in some way? Or would it be better to excute scripts-20070109-1725/training/symal/giza2bal.pl (what is the command for this?) >From what I can see, the last thing it tried to execute was: bin/moses-scripts/scripts-20080811-1801/training/phrase-extract/extract /Users/lliohumphreys/MT/MOSESSUITE/model/aligned.0.en /Users/lliohumphreys/MT/MOSESSUITE/model/aligned.0.it /Users/lliohumphreys/MT/MOSESSUITE/model/aligned.grow-diag-final-and /Users/lliohumphreys/MT/MOSESSUITE/model/extract.0-0 7 orientation Maybe I should amend and run this command? If successful, would that complete the following step in the baseline model: Run training script: bin/moses-scripts/scripts-YYYYMMDD-HHMM/training/train-factored-phrase-model.perl -scripts-root-dir bin/moses-scripts/scripts-YYYYMMDD-HHMM -root-dir working-dir -corpus working-dir/corpus/europarl.lowercased -f fr -e en -alignment grow-diag-final-and -reordering msd-bidirectional-fe -lm 0:5:working-dir/lm/europarl.lm:0 Many thanks, Llio ---------- Forwarded message ---------- From: Llio Humphreys <[EMAIL PROTECTED]> Date: Aug 12, 2008 2:53 PM Subject: Fwd: [Moses-support] Error in training phrase To: [EMAIL PROTECTED] ---------- Forwarded message ---------- From: Philipp Koehn <[EMAIL PROTECTED]> Date: Wed, Nov 28, 2007 at 4:03 PM Subject: Re: [Moses-support] Error in training phrase To: Nguyen Tien Huy <[EMAIL PROTECTED]> Cc: "[email protected]" <[email protected]> Hi, either the GIZA++ runs or the merging of their output with with the word alignment program "symal" failed. Check the output of the GIZA++ runs, it should look something like this: -rw-r--r-- 1 pkoehn people 1694973594 Nov 23 07:26 eng-chi.A3.final.gz -rw-r--r-- 1 pkoehn people 2117713098 Nov 13 16:33 eng-chi.cooc -rw-r--r-- 1 pkoehn people 2161 Nov 13 16:34 eng-chi.gizacfg If that is OK, check, if you can run the symal command by hand. (check for "Executing: /.../scripts-20070109-1725/training/symal/giza2bal.pl") in the STDERR of the training script. -phi On Nov 28, 2007 11:38 AM, Nguyen Tien Huy <[EMAIL PROTECTED]> wrote: > > Dear all, > > Below is contents of model directory: > > ------------------------------------------------------------------------- > [EMAIL PROTECTED]:~/Project/testdata/model$ ls -l > total 2632 > -rw-r--r-- 1 nguyen root 815863 2007-11-27 02:47 aligned.0.en > -rw-r--r-- 1 nguyen root 924249 2007-11-27 02:47 aligned.0.fr > -rw-r--r-- 1 nguyen root 0 2007-11-27 02:47 aligned.grow-diag-final-and > -rw-r--r-- 1 nguyen root 0 2007-11-27 02:48 extract.0-0.o > -rw-r--r-- 1 nguyen root 463493 2007-11-27 02:48 lex.0-0.f2n > -rw-r--r-- 1 nguyen root 463493 2007-11-27 02:48 lex.0-0.n2f > ---------------------------------------------------------------------------- > Size of files are small because I only used 10000 sentences from original > files. > > Thanks > > ________________________________ > > Date: Tue, 27 Nov 2007 12:56:12 +0000 > > From: [EMAIL PROTECTED] ed.ac.uk > > To: [EMAIL PROTECTED] > > > Subject: Re: [Moses-support] Error in training phrase > > CC: [email protected]; [EMAIL PROTECTED]; [EMAIL PROTECTED] > > > > > Hi, > > > > can you do a "ls -l model" and email the contents of that directory? > > > > -phi > > > > On Nov 27, 2007 2:14 AM, Nguyen Tien Huy <[EMAIL PROTECTED]> wrote: > > > > > > Dear all, > > > > > > Thanks for your response. > > > > > > I've passed this error by change param "-CoocurrenceFile" to > > > "-coocurrencefile", but no luck on step 5 (extract phrases) > > > > > > > > > > ------------------------------------------------------------------------------------------------------------------------------------------------- > > > [EMAIL PROTECTED]:~$ perl > > > $SCRIPTS_ROOTDIR/training/train-factored-phrase-model.perl > -scripts-root-dir > > > $SCRIPTS_ROOTDI! R -root-dir /home/nguyen/Project/testdata -corpus > > > /home/n guyen/Project/testdata/corpus/prototype.lowercased -f fr -e en > > > > -alignment grow-diag-final-and -reordering msd-bidirectional-fe -lm > > > 0:5:/home/nguyen/Project/testdata/lm/prototype.en.lm:0 > > > ......... > > > Saved: /home/nguyen/Project/testdata/model/lex.0-0.f2n and > > > /home/nguyen/Project/testdata/model/lex.0-0.n2f > > > (5) extract phrases @ Tue Nov 27 02:48:36 ICT 2007 > > > (1.0.5) reducing factors to produce /home/nguyen/Project/te! > > > stdata/model/aligned.0.fr @ Tue Nov 27 02:48:36 ICT 2007 > > > /home/nguyen/Project/testdata/model/aligned.0.fr in place, reusing > > > (1.0.5) reducing factors to produce > > > /home/nguyen/Project/testdata/model/aligned.0.en @ Tue Nov 27 02:48:36 > ICT > > > 2007 > > > /home/nguyen/Project/testdata/model/aligned.0.en in place, reusing > > > (5) [0-0] extract phrases @ Tue Nov 27 02:48:36 ICT 2007 > > > /home/ngu! > yen/Project/bin/moses-scripts/scripts-20071126-0027/training/phrase-extract/extract > > > > /home/nguyen/Project/testdata/model/aligned.0.en > > > /home/nguyen/Project/testdata/model/aligned.0.fr > > > /home/nguyen/Project/testdata/model/aligned.grow-diag-final-and > > > /home/nguyen/Project/testdata/model/extract.0-0 7 orientation > > > Executing: > > > > /home/nguyen/Project/bin/moses-scripts/scripts-20071126-0027/training/phrase-extract/extract > > > /home/nguyen/Project/testdata/model/aligned.0.en > > > /home/nguyen/Project/testdata/model/aligned.0.fr > > > /home/nguyen/Project/testdata/model/aligned.grow-diag-final-a! nd > > > /home/nguyen/Project/testdata/model/extract.0-0 7 orientation > > > Ph raseExtract v1.3.0, written by Philipp Koehn > > > phrase extraction from an aligned parallel corpus > > > (also extracting orientation) > > > Executing: cat /home/nguyen/Project/test! data/model/extract.0-0.o.part* > > > > > > > /home/nguyen/Project/ testdata/model/extract.0-0.o > > > cat: /home/nguyen/Project/testdata/model/extract.0-0.o.part*: No such > file > > > or directory > > > Exit code: 1 > > > Died at > > > > /home/nguyen/Project/bin/moses-scripts/scripts-20071126-0027/training/train-factored-phrase-model.perl > > > line 879. > > > > ------------------------------------------------------------------------------------------------------------------------------------------------- > > > Could you please point me to a solution? > > > Thank in advance, > > > Huy. > > > > > > ________________________________ > > > > Date: Mon, 26 Nov 2007 17:25:22 +0100 > > > > From: [EMAIL PROTECTED] > > > > > > > To: [email protected] > > > > Subject: Re: [Moses-support] Error in training phrase > > > > > > > > > > > Hi, > > > > > > > > Philipp Koehn a écrit :! > > > >! > > > > > you need to compile GIZA++ with the option > -DBINARY_SEARCH_FOR_TTABLE > > > > > > > > Possible but not necessaray. I found that compiling with > > > > -DBINARY_SEARCH_FOR_TTABLE leads to a slower training. I would use > > > > -DBINARY_SEARCH_FOR_TTABLE only when the training data is so large > that > > > > GIZA++ runs out of memory. > > > > > > > > > On Nov 22, 2007 4:10 PM, Nguyen Tien Huy <[EMAIL PROTECTED]> > wrote: > > > > > > > > > > > > ERROR: parameter 'coocurrencefile' does not exist. > > > > > > WARNING: ignoring unrecognized option: -CoocurrenceFile > > > > > > ERROR: parameter 'prototypegizavienviencooc' does not exist. > > > > > > WARNING: ignoring unrecognized option: > > > > > > prototype/giza.vi-en/vi-en.cooc ..... > > > > > > ERR! OR: parameter 'coocurrencefile' does not exist. > > > > > > > WARNING: ignoring unrecognized option: -CoocurrenceFile > > > > > > ERROR: parameter 'protot! ypegizaenvienvicooc' does not exist. > > > > > > WARNING: ignori ng unrecognized option: > > > > > > > > > prototype/giza.en-vi/en-vi.cooc ..... > > > > > > > > Those should be warnings, not errors, and are harmless. > > > > > > > > > > ..Executing: cat prototype/model/extract.0-0.o.part* > > > > > > > prototype/model/extract.0-0.o > > > > > > cat: prototype/model/extract.0-0.o.part*: No such file or > directory > > > > > > > > This, is the problem. Any previous error messages? (For instance, disk > > > > full?) > > > > > > > > -- Daniel > > > > _______________________________________________ > > > > Moses-support mailing list > > > > [email protected] > > &g! t; > http://mailman.mit.edu/mailman/listinfo/moses-support > > > > > > > ________________________________ > > > Get news, entertainment and everything you care about at Live.com. Check > it > > > out! > > ________________________________ > Connect to the next generation of MSN Messenger Get it now! _______________________________________________ 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
