The Maximum Entropy Toolkit provides a set of tools and library for constructing maximum entropy (maxent) model in either Python or C++.
Maxent Entropy Model is a general purpose machine learning framework that has
proved to be highly expressive and powerful in statistical natural language
processing, statistical physics, computer vision and many other fields. Please
see my maxent page for more information on maxent.
right place.
If you do not like this Toolkit, you can still try several other ME implementations.
Current version has the following highlights:
The toolkit is written in ISO C++ with speed and portability in mind. It has been tested under the following OS/Compilers:
This software is freeware and is released under LGPL license. Please consult the LICENSE file in source package for more information.
The adoption of LGPL is in accord with the license of java maxent project: http://maxent.sourceforge.net, from which the toolkit was derived. LGPL makes it easier to share source code, as well as new ideas, between both projects.
Here's a list of files you can download. Please note:
File Name | MD5 Sum | Description |
Source Code | ||
maxent-20041229.tar.gz | b9181d6f2ed97815fa6246f43d8c5464 | The source code of the latest version plus document |
Various Binaries | ||
maxent-20041229-linux-static.tar.gz | de12f65e8935c2c81aaca0c0c628377d | Statically linked binary for Linux |
maxent-20041229-mingw32-static.zip | 1f4e8a49f226f16b99e27d9db0b08534 | Statically linked win32 binary cross compiled from a Linux box |
maxent-20041229.win32-py2.3.exe | 3201e53e5bbe3a7df53d25bd93738865 | Win32 Python23 extension setup program (built with MSVC7.1 + Intel Fortran 8.0) |
maxent-20041229.win32.zip | 487c5e69053b5f6a51c85f55f9cf0869 | Win32 Python23 extension package (just extract all files to your python directory if you failed to run the above one) |
Other Files | ||
jam.exe | 9992097c5ead36f3dd54f34e8e20302e | Win32 Jam binary, you need this in order to build the software on win32 with compilers other than GCC |
tagger.tar.gz | 3dea4c7c2683e6021d89a9d7c9932c5c | (Optional) The example POS tagging model that was trained on WSJ 00-18 sections is now downloaded separately (see document for more information) |
Here is the PDF manual. Alternatively, you can generate HTML reference document by running "doxygen" in the doc/ directory if you have Doxygen installed.
IFLAG= -1 LINE SEARCH FAILED. SEE DOCUMENTATION OF ROUTINE MCSRCH ERROR RETURN OF LINE SEARCH: INFO= 3 POSSIBLE CAUSES: FUNCTION OR GRADIENT ARE INCORRECT OR INCORRECT TOLERANCES std::runtime_error caught:lbfgs routine stops with an error
This is a rare event and most users do not encounter this problem. However, when this does happen it is usually an indication of "over-training", which can be caused by either using a very small dataset, or trying to perform too many training iterations on a large dataset. The ad-hoc solution is to use a larger dataset, or simply use fewer training iterations. Future releases will probably turn this bug into a useful feature:-)
You will get this error if you try to compile the source code with an old version of gcc (gcc 2.9x). This can be fixed by using gcc 3.x instead. If your new gcc binaries are named gcc32 and g++32. You can run the following command to get the building system configured properly:
CC=gcc32 CXX=g++32 ./configure
There is no document on the model file format at this time. Since it is likely to be changed in future releases. You are encouraged to look at src/modelfile.cpp, if you are really curious.
Back to the year 2002, as I got more knowledge on maxent I decided to do some experiments to assess the usefulness of the framework. Unfortunately, unlike other popular machine learning techniques such as SVMs, which have many off-the-shelf implementations on the net, only few maxent packages can be found from the internet. Partly because implementing a maxent model is a non-trival work, especially the iterative parameter estimating procedure.
After trying several maxent implementations, I found the java maxent package is a high quality, easy use one. The software is part of the OpenNLP project and is written in Java. But I want a C++ or Python solution that can be integrated into my existing code. Then, I began rewriting the java code in C++. Hopefully, it is relatively easy to translate java code into C++ (but not vice versa) and by the end of 2002 I finished the first C++ version. The speed of C++ is impressive: optimized C++ code ourperforms java code by a wide margin.
In the spring of 2003, python binding was added, utilizing Boost.Python lib. Later I found Dr. Malouf's paper, which proposes to use Limited Memory BFGS Method to estimate ME model's parameters. His experiment showed L-BFGS was much faster than GIS and IIS. So, I added L-BFGS estimating code in May. 2003. Meanwhile, I swithed the project form autoconf/automake to SCons, a much better make replacement written in Python. Later, in April. 2003 I came across Curran and Clark's paper, which proves a correction-free version of GIS algorithm (originally pointed out by [Goodman, 2002]), elimiting the need of correction feature in GIS. I readily adopted the idea and greatly simplified the GIS code.
In September, 2003, documentation was added and the toolkit was put to public release.
If you have any questions or comments regarding the use of the software, the
application of MaxEnt technique, or simply want to have a (related)
discussion with me, feel free to drop me a line:
The Author: Zhang Le
<
ejoy@users.sourceforge.net>
Last Change :12-Apr-2005. Please send any question to Zhang Le |