Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Maximum Entropy Modeling Toolkit for Python and C++
C++ Python C FORTRAN Other
Tree: 71c5e6a67b

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.
doc
example
lib/boost
python
script
src
test
tools
ChangeLog
INSTALL
Jamfile
Jamrules.in
LICENSE
Makefile
README
THANKS
aclocal.m4
config.h.in
configure
configure.in

README

A Maximum Entropy Modeling Toolkit for Python and C++
=====================================================


http://homepages.inf.ed.ac.uk/lzhang10/maxent.html

What is this?
-------------
This package provides a (Conditional) Maximum Entropy Modeling Toolkit for
Python and C++. The library is written in ISO C++ and has been tested under
GNU/Linux, FreeBSD/NetBSD and Win32 (Cygwin and http://www.mingw.org[MinGW]).
There is also a python extension module (maxent module) available.

For more information, please have a look at the file manual.pdf in doc/
directory. The PDF document talks about the toolkit at length.

Building
--------
See file INSTALL for detail description of building the maxent package on unix
platforms. Instruction on building under win32 environment is covered in the
PDF manual in doc/.

python/ directory contains source code for python binding of the toolkit. The
detail instruction on building the python extension is given in python/README.

License
-------
This software is freeware and is released under LGPL. see LICENSE file for
more information. 

The adoption of LGPL is in accord with the license of java maxent project:
http://maxent.sf.net, from which the toolkit was derived.  LGPL makes it
easier to share source code, as well as new ideas, between both projects.

Misc Info
---------
About Feature Value
~~~~~~~~~~~~~~~~~~~
Since binary feature and integer feature are more common than high precision
real feature, all feature values are represented as float (4 bytes) rather
than double (8 bytes) in order to save memory. If you need to specify double
feature values you can find all floats in source code and replace them with
double (require more memory).

Undocumented Functions
~~~~~~~~~~~~~~~~~~~~~~
There are some functions in the source code that are not documented. They are
considered experimental and are very likely to be changed in future releases. 
Please do not report them as bugs.

Contact Author
--------------
Le Zhang <mailto:ejoy@users.sf.net[]> 
http://homepages.inf.ed.ac.uk/lzhang10/
Question and suggestion are always welcome.
Something went wrong with that request. Please try again.