On 22/12/2016 05:26, WULongski wrote:
Hi,
I have two questions:
1、
After reading the adding feature section of the
Moses(http://www.statmt.org/moses/?n=Moses.FeatureFunctions), it says
"However, new features are usually more complicated. For instance, it
requires reading in a file and representing it with a data structure
and more complex computations. See moses/LM/SRI.h and
moses/LM/SRI.cpp for something more involved."
while my moses.ini is looking as this:
//////////////////////// moses.ini
UnknownWordPenalty
WordPenalty
PhrasePenalty
PhraseDictionaryCompact name=TranslationModel0 num-features=4
path=/nobackup/f1/asr/wulong/Moses/mosesdecoder/ALLDATApinyin/working/filtered-oneword/phrase-table.0-0.1.1
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=/nobackup/f1/asr/wulong/Moses/mosesdecoder/ALLDATApinyin/working/filtered-oneword/reordering-table.wbe-msd-bidirectional-fe.0-0.1
Distortion
KENLM name=LM0 factor=0
path=/nobackup/f1/asr/wulong/Moses/mosesdecoder/ALLDATApinyin/lm/train.blm.en
order=3
//////////////////////////
the last line is the KENLM , this means that language model
estimation is did by the KENLM. But I installed the moses by the way
on the Mose website: " Easy Setup on Ubuntu (on other linux systems,
you'll need to install packages that provide gcc, make, git, automake,
libtool) " in the
_http://www.statmt.org/moses/?n=Development.GetStarted_.
But in the third step of the installing process, it says " 3. Run
the following to install a recent version of Boost (the default
version on your system might be too old), as well as cmph (for
CompactPT), irstlm (language model from FBK, required to pass the
regression tests), and xmlrpc-c (for moses server). By default, these
will be installed in ./opt in your working directory:make -f
contrib/Makefiles/install-dependencies.gmake ". It means that the
language model estimation is did by the irstlm .
no - it means IRSTLM available be used in the decoder, but it doesn't
have to be used. KENLM is always available
so I am confused why is KENLM in the moses.ini.
2、
"KENLM name=LM0 factor=0
path=/nobackup/f1/asr/wulong/Moses/mosesdecoder/ALLDATApinyin/lm/train.blm.en
order=3"
can you explain the parameters of the above line? i know order =3
means 3-gram , path is the arpa table language model .But i do not
know the other parameter meaning.
name - the unique name. Must match the name in the [weight] section
factor - which factor the LM will use. 0 if you are not doing factored
models
Thank you very much .
_______________________________________________
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