|
Downloading and Building SRILM
Building SRILM
SRILM is available only in source form.
We cannot offer precompiled binaries, so you will have to download
the source code and build and install the software yourself.
Instructions can be found in the INSTALL file
in the top-level distribution directory.
If you have previously downloaded SRILM you
may want to check for the current release version
and the change log.
Prerequisites
Besides the SRILM distribution,
you will also need the following freely available tools:
-
A template-capable ANSI-C/C++ compiler, preferably
GCC version 3.4.3 or higher.
Be sure to build or download g++ (the C++ compiler) component of GCC.
Recent versions of
Microsoft Visual Studio Express
should also work.
-
GNU make,
to control compilation and installation.
-
GNU gawk,
required for many of the utility scripts.
-
GNU gzip
to unpack the distribution, and to allow SRILM programs to
handle ".Z" and ".gz" compressed datafiles (highly recommended).
-
bzip2
to handle ".bz2" compressed files (optional).
-
p7zip
to handle "7-zip" compressed files (optional).
-
XZ Utils
to handle ".xz" compressed files (optional).
-
The iconv library
or equivalent, to handle UTF-encoded text files on Unix-based systems.
(This library may already be integrated in your C library.)
-
The Tcl
embeddable scripting language library (only required for
some of the test executables).
-
The Cygwin
porting layer, to build SRILM on a Microsoft Windows system.
Cygwin comes with the packages above,
although they may not all be installed by default.
Alternatively, you can build native Windows binaries using the free
MS Visual Studio 2010 Express
compiler.
Noncommercial Use
Government agencies, and schools, universities, and non-profit organizations
can download
SRILM free of charge under
SRI's "Research
Community License", for use in projects that do not receive
external funding other than government research grants and contracts.
For other uses please inquire about
commercial licensing.
Download
To download the source code, please first fill in the form below.
Read the License document and indicate
your agreement to comply with it by pressing the button below.
This will start the download.
The distribution is in gzipped tar format, so you should tell your browser to
store the data in a file ending in ".tar.gz".
You will need the "gunzip" and "tar" utilities to unpack the downloaded
file into a directory structure.
The only use we will make of your address information is as a record
of the licensing and to inform you of software updates.
It will not be passed on to third parties.
If you decide the License or the software is not for you, please do not
proceed to download the software.
If you do want to use SRILM or are generally interested in it, please consider
joining the SRILM user mailing list.
References
Published research using SRILM may cite the following paper:
A. Stolcke (2002),
SRILM -- An Extensible Language Modeling Toolkit.
Proc. Intl. Conf. on Spoken Language Processing,
vol. 2, pp. 901-904, Denver.
Research making use of factored language models should also cite:
J. A. Bilmes and K. Kirchhoff (2003),
Factored language models and generalized parallel backoff,
Proc. HLT-NAACL, pp. 7-9, Edmonton, Alberta.
Research making use of the maximum entropy modeling functionality should also cite:
T. Alumäe and M. Kurimo (2010),
Efficient Estimation of Maximum Entropy Language Models with N-gram features: an SRILM extension,
Proc. Interspeech, pp. 1820-1823, Makuhari, Japan.
Back to SRILM main page.
|
|