Hi,

so you are spending a lot of time on start-up. There is various
things you could do about it. One is to use binarized translation
tables and IRSTLM, which have shorter start-up time.

But then, you may not care too much about the start-up time,
if you keep the decoder running and feed it sentences without
restarting. This is what is done in the online web translation
set-up, so you may want to look at that.

-phi

On Tue, May 5, 2009 at 7:01 PM, Jan Helak <yac...@poczta.onet.pl> wrote:
> Hello,
>> Hi,
>>
>> 4 minutes sounds like too much for translating a sentence.
>> A little bit more detail would be helpful to detect where the
>> time savings could be:
>> - does the time include the start-up time (loading translation
>>   and language models)?
>> - do you use IRSTLM, SRILM, binarized phrase and reordering tables?
>>
> I just went through a tutorial http://www.statmt.org/moses_steps.html
> into end of Part II.
> I used SRILM to build a language model.
>
>
> More information about time:
> Finished loading LanguageModels : [46.000] seconds
> Finished loading phrase tables : [213.000] seconds
> End. : [213.000] seconds
>
> Again I add a listing with commands used to create decoder:
> *******************
>
> cat data/training/beta.en | tools/scripts/tokenizer.perl -l en >
> work/corpus/beta.tok.en
>
> cat data/training/beta.pl | tools/scripts/tokenizer.perl -l en >
> work/corpus/beta.tok.pl
>
> tools/moses-scripts/scripts-20090321-2154/training/clean-corpus-n.perl
> work/corpus/beta.tok pl en work/corpus/beta.clean 1 40
>
> tools/scripts/lowercase.perl < work/corpus/beta.clean.en 
> >work/corpus/beta.lowercased.en
> tools/scripts/lowercase.perl < work/corpus/beta.clean.pl 
> >work/corpus/beta.lowercased.pl
>
>
> tools/scripts/lowercase.perl < work/corpus/beta.tok.en > 
> work/lm/beta.lowercased.en
>
> tools/srilm/bin/i686/ngram-count -order 3 -interpolate -kndiscount -unk -text 
> work/lm/beta.lowercased.en -lm work/lm/beta.lm
>
> tools/moses-scripts/scripts-20090321-2154/ -root-dir work -corpus
> work/corpus/beta.lowercased -f pl -e en -alignment grow-diag-final-and
> -reordering msd-bidirectional-fe -lm 0:3:/home/janek/Lik/work/lm/beta.lm >& 
> work/training.out &
>
> ********************************************************************
>
>
> regards,
> JH
> UAM
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>

_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to