This library is outdated. We are working on new stuff currently in pylearn2 here: https://github.com/lisa-lab/pylearn2/

Pylearn is a Python library for machine learning, built on top of Theano, our library for defining, optimizing and evaluating mathematical expressions involving multi-dimensional arrays.

This documentation is under construction, but you can already access the automatically-generated API doc, along with more extensive explanations for some modules.


We recommend the latest development version, available via:

.. code-block:: bash

The Pylearn/pylearn subfolder contains the Python package and must be made available to Python programs, for instance by adding the Pylearn root folder to your PYTHONPATH environment varialble:

export PYTHONPATH=$PYTHONPATH:/path/to/Pylearn

Assembla project: https://www.assembla.com/spaces/pylearn/


For the moment, the following documentation is available.

  • Formulas – Built-in math formulas optimized for speed and robustness
  • io.SeriesTables module – Saves error series and other statistics during training
  • API – The automatically-generated API documentation
  • V2 planning – Some documentation about the planning of our next version of pylearn.

You can download the latest PDF documentation, rather than reading it online.