IRSTLM

IRSTLM is a free and open source exact statistical language model using memory-mapping. The language models are compatible with those created with the closed-source SRILM Tooolkit.

See the homepage at http://hlt.fbk.eu/en/irstlm

Installation

svn checkout svn://svn.code.sf.net/p/irstlm/code/trunk irstlm
cd irstlm
cmake -G "Unix Makefiles" -DCMAKE_INSTALL_PREFIX=/path/prefix
make -j4
make install

Make a language model

export IRSTLM=/prefix/you/used/when/installing/irstlm
$IRSTLM/build-lm.sh -i incorpus.txt -o out.lm.gz -t tmp/

See also

• Moses
• Using GIZA++
• RandLM - a randomised LM, based on Bloom Filters