How to install Moses (Statistical Machine Translation) on Ubuntu?

In this article, I will show you how to install and build Moses on Ubuntu, and how to use Moses to translate with some simple models (English and Sign Language Gloss). If you experience problems, then please contact me.

If you’re just writing about this work, please cite this paper as follow:

Achraf Othman, Mohamed Jemni, “Designing High Accuracy Statistical Machine Translation for Sign Language Using Parallel Corpus—Case study English and American Sign Language “, Journal of Information Technology Research, Volume 12, Issue 2, 2019.

Part 1 – Installing Moses and Giza++

1. Login as a root
2. Navigate to user directory
3. Create a workplace
   mkdir smt
   cd smt
4. Install required packages
   sudo apt-get install build-essential git-core pkg-config automake libtool wget zlib1g-dev python-dev libbz2-dev
5. Install required packages for regression step
   sudo apt-get install libsoap-lite-perl
6. Clone Moses from Github
   git clone https://github.com/moses-smt/mosesdecoder.git
7. Download & Install GIZA++
   git clone https://github.com/moses-smt/giza-pp.git
   cd giza-pp
   make
8. Copying GIZA++ Binaries to MosesDecoder
   cd ../mosesdecoder
   mkdir tools
   cp ../giza-pp/GIZA++-v2/GIZA++ ../giza-pp/GIZA++-v2/snt2cooc.out ../giza-pp/mkcls-v2/mkcls tools
   cd ..

Part 2 – Installing IRSTLM

1. Install IRSTLM 5.80.08 (Download & Extract IRSTLM from SourceForge, to your home directory)
   mkdir irstlm
   cd irstlm-5.80.08
   cd trunk
   ./regenerate-makefiles.sh
   ./configure –prefix=$HOME/irstlm [$HOME = the path where you want to install irstlm]
   make install
   cd ..
   cd ..

if you are facing challenging to make files like “cstdlib – fatal error: stdlib.h: No such file or directory“, please watch the video to see how to downgrade the version of gcc and g++ in order to build IRSTLM 5.80.08 successfully.

Part 3 – Installing Boost, CMPH, and, XML-RPC

1. Install Boost 1.64  (This may take a lot of time! be patient 😉 )
   wget https://dl.bintray.com/boostorg/release/1.72.0/source/boost_1_72_0.tar.gz
   tar zxvf boost_1_72_0.tar.gz
   cd boost_1_72_0/
   ./bootstrap.sh
   ./b2  –layout=system link=static install || echo FAILURE
   cd ..
   
2. install CMPH2.0
   wget http://www.achrafothman.net/aslsmt/tools/cmph_2.0.orig.tar.gz
   tar zxvf cmph_2.0.orig.tar.gz
   cd cmph-2.0/
   ./configure
   make
   make install
3. install XML-RPC
   wget http://www.achrafothman.net/aslsmt/tools/xmlrpc-c_1.33.14.orig.tar.gz
   tar zxvf xmlrpc-c_1.33.14.orig.tar.gz
   cd xmlrpc-c-1.33.14/
   ./configure
   make
   make install
   cd ..
   
4. Install Moses (This may take a lot of time! 😯 )
   cd mosesdecoder
   make -f contrib/Makefiles/install-dependencies.gmake
   ./bjam –with-boost=../boost_1_72_0 –with-cmph=../cmph-2.0 –with-irstlm=../irstlm 

The next step is to train models and test our statistical machine translation.

If you’re just writing about this work, please cite this paper as follow:

Achraf Othman, Mohamed Jemni, “Designing High Accuracy Statistical Machine Translation for Sign Language Using Parallel Corpus—Case study English and American Sign Language “, Journal of Information Technology Research, Volume 12, Issue 2, 2019.