thanks. It may not be the phrase-based/hiero mixup as I had suggested. Maybe the system ran out of RAM.
Before you run tuning, you should also binarize the phrase-table and lexicalised reordering model, this can be done in the filtering process by providing the absolute path of the binarizer, eg -Binarizer "/home/s0565741/workspace/github/hh/bin/processPhraseTableMin This shows you an example http://www.statmt.org/moses/RELEASE-3.0/models/de-en/steps/1/TUNING_filter.1 Binarizing reduces RAM usage by only loading translation rules for the particular sentence it is translating * Looking for MT/NLP opportunities * Hieu Hoang http://moses-smt.org/ On 17 May 2017 at 11:42, Per Starbäck <starb...@stp.lingfil.uu.se> wrote: > Hieu Hoang writes: > > It may be you're mixing up the phrase-based with the syntax/hiero model. > Can you please > > attach the file > > filtered/moses.ini > > and perhaps the first few lines of the phrase-table that's used > > Sure! Here is mert-work/filtered/moses.ini, again just with the > beginning of absolute paths replaced with "...": > > ====================================================================== > # input factors > [input-factors] > 0 > > # mapping steps > [mapping] > 0 T 0 > > [distortion-limit] > 6 > > # feature functions > [feature] > UnknownWordPenalty > WordPenalty > PhrasePenalty > PhraseDictionaryMemory name=TranslationModel0 num-features=4 > path=.../mert-work/filtered/phrase-table.0-0.1.1.gz input-factor=0 > output-factor=0 > LexicalReordering name=LexicalReordering0 num-features=6 > type=wbe-msd-bidirectional-fe-allff input-factor=0 output-factor=0 > path=.../mert-work/filtered/reordering-table.wbe-msd-bidirectional-fe > Distortion > KENLM lazyken=0 name=LM0 factor=0 path=.../lm/news-commentrary-v8.de-en.blm.en > order=3 > > # dense weights for feature functions > [weight] > UnknownWordPenalty0= 1 > WordPenalty0= -1 > PhrasePenalty0= 0.2 > TranslationModel0= 0.2 0.2 0.2 0.2 > LexicalReordering0= 0.3 0.3 0.3 0.3 0.3 0.3 > Distortion0= 0.3 > LM0= 0.5 > ====================================================================== > > And here the beginning of mert-work/filtered/phrase-table.0-0.1.1.gz > uncompressed: > > ====================================================================== > $ ||| $100 they earn. ||| 0.25 0.0252101 0.025 1.54085e-10 ||| 0-0 ||| 4 > 40 1 ||| ||| > $ ||| $100 they ||| 0.25 0.0252101 0.025 2.56808e-05 ||| 0-0 ||| 4 40 1 > ||| ||| > $ ||| $100 ||| 0.00819672 0.0252101 0.025 0.0084507 ||| 0-0 ||| 122 40 1 > ||| ||| > $ ||| $100,000 apiece in campaign contributions in ||| 0.5 1 0.025 > 1.16584e-20 ||| 0-1 ||| 2 40 1 ||| ||| > $ ||| $100,000 apiece in campaign contributions ||| 0.5 1 0.025 > 6.46129e-19 ||| 0-1 ||| 2 40 1 ||| ||| > $ ||| $100,000 apiece in campaign ||| 0.5 1 0.025 1.54207e-14 ||| 0-1 ||| > 2 40 1 ||| ||| > ====================================================================== >
_______________________________________________ Moses-support mailing list Moses-support@mit.edu http://mailman.mit.edu/mailman/listinfo/moses-support
