The CMU Sphinx Group Open Source Speech Recognition Engines

Speech at CMU   |   Sphinx at SourceForge

Introduction

General Documentation

CMUSphinx Components

Common library

Decoders

Acoustic Model Training

Language Model Training

Utilities


Latest News

Sphinx4-1.0beta3 released
2009-08-17 19:23
Read More »

SphinxTrain 1.0 Released
2009-02-12 16:05
Read More »

Sphinx-4 1.0 beta2 released
2009-02-07 18:27
Read More »

Site news archive »


External Links

Notice: if you have comments about the links below, please contact the authors directly.

CMU Sphinx documentation Wiki

This page details the things you need to install on Windows, Linux, and Mac OS X in order to do Python development with SphinxTrain.

Note that Python is not currently required to use SphinxTrain, unless you wish to use LDA and MLLT feature transforms. However, prototyping and research is currently being done primarily using Python.

First, install Python. For Windows and Mac OS X, the latest release can be found at http://www.python.org/download/. On Linux, it should be installed with your distribution.

Next, install NumPy and SciPy. For Windows, there are binary packages at http://scipy.org/Download. On Mac OS X, you may be able to find it via Fink or DarwinPorts. On Linux, it should be available in most recent distributions. On Ubuntu and Debian, for instance, you can just run:

apt-get install python-scipy

If you don't have a binary package to install, then you can compile it from source following the instructions on http://scipy.org/Download.

You can also optionally install iPython and matplotlib for enhanced interactive use and MATLAB-like graphics.

Now, if you wish to install the SphinxTrain Python modules system, wide, go to the python directory inside your SphinxTrain installation and run:

python setup.py build
sudo python setup.py install

InstallingPythonStuff (last edited 2008-02-21 03:34:54 by localhost)

SourceForge.net Logo This page is maintained by David Huggins-Daines ()
CMUSphinx is a project within the Sphinx Group at Carnegie Mellon