Hi,
I am using experimenter to run Moses. In the configuration file, I provide
word alignments by specifying it as"word-alignment =" in TRAINING. I could
not find a way to specify cleaned corpus. I am using the latest branch of
Moses. In the previous one, I used to specify cleaned corpus in the
TR
I'm doing some LM tinkering, and I noticed that there's a new parameter in
FeatureFunction and its descendents.
What is the following parameter represent?
"const std::string &line"
Thanks,
Lane
___
Moses-support mailing list
Moses-support@mit.edu
http:
Hi,
Might be two reasons:
1) are you running this on a 32-bit machine or Cygwin? Then the maximum
size phrase table that can be built is about 3GB, which can be hit quite
quickly.
2) do you have only little free space in your /tmp directory? You can
change the directory used for temporary fil
Hello,
I am trying to create a compact version of the phrase table based on the
pre-trained models of release 1.
I get the following error when I run the processPhraseTableMin on a ubuntu
Vagrant machine.
Thanks for your help,
João
vagrant@precise64:/vagrant/mt-models/en-es$
~/mosesdecoder/bin
did you escape your training and input data? There must not be |
characters in your data unless you are using factored models
the moses tokenizer script does it, as well as the specific escape script.
scripts/tokenizer/tokenizer.perl
scripts/tokenizer/escape-special-chars.perl
On 26/08/2
you mean your lattice only has 1 path? ie. it is a sentence encoded as a
lattice?
Did you retune your model?
I am surprise, but no-one has done that experiment. if you can make the
model files available, i can debug it and let you know why there is a
difference
On 26/08/2013 14:42, Wei Qiu
Hi Nadir,
Thanks a million, it really helped!
After switching e and f positions at the command line, it worked fine on my
test corpus. Also worked on the new Perl script.
Thanks again,
Jian
On Tue, Aug 27, 2013 at 1:50 PM, Nadir Durrani wrote:
> Hi Jian,
>
> Try out
>
> ${MOSES_DIR}/script
Hi Hieu and Phillip,
Thanks a lot for your help.
Hieu, sorry for the q problem. I was tired, should have payed more
attention.
Best regards,
João Graça
On Tue, Aug 27, 2013 at 2:26 PM, Hieu Hoang wrote:
> internally, the decoder also does this conversion.
>
> however, it might be better t
internally, the decoder also does this conversion.
however, it might be better to use philipp's script to be more transparent
On 27 August 2013 13:58, Philipp Koehn wrote:
> Hi,
>
> note that the script
> scripts/training/convert-moses-ini-to-v2.perl < IN > OUT
> allows you to easily convert
Hi,
note that the script
scripts/training/convert-moses-ini-to-v2.perl < IN > OUT
allows you to easily convert an old config file to the new format.
-phi
On Tue, Aug 27, 2013 at 1:47 PM, Hieu Hoang wrote:
> there's a character
>q
> at the beginning of the file.
>
> Delete it
>
>
> On 26 A
Hi Jian,
Try out
${MOSES_DIR}/scripts/OSM/OSM-Train.sh ${TARGET} ${SOURCE} ${ALIGN} ${ORDER}
${WORKDIR} ${MOSES_DIR} ${SRILM_BINDIR}
the script takes target before source. Btw we made some changes. There's a
new script OSM-Train.perl. So use that now.
I have also uploaded the source code for ge
OK, I was hoping for a dirty method to bypass that. No pressure :)
W dniu 2013-08-27 14:38, Kenneth Heafield napisał(a):
> Hi,
>
> I look forward to working with you at MT Marathon on adding features to
> lmplz.
>
> Kneser-Ney smoothing is not well defined when there is no singleton.
> You
there's a character
q
at the beginning of the file.
Delete it
On 26 August 2013 22:34, João Graça wrote:
> Dear Hieu,
>
> I am trying to use the new version of moses with the old format moses.ini
> from the pre-trained models.
>
> In attach is the configuration file I am currently trying to
Hi,
I look forward to working with you at MT Marathon on adding features to
lmplz.
Kneser-Ney smoothing is not well defined when there is no singleton.
You'll get a similar error message from SRILM on this data with KN. I
suggest you consider a different smoothing method. Curren
Hi Kenneth,
I am getting the following error from lmplz on training data built from
POS-tags:
/home/m.junczys/kenlm/lm/builder/adjust_counts.cc:42 in void
lm::builder::{anonymous}::StatCollector::CalculateDiscounts() threw
BadDiscountException because `s.n[j] == 0'.
Could not calculate Knese
15 matches
Mail list logo