This directory contains examples of how to use pylearn2 DBMs

1. How to train a single layer DBM, aka an RBM

    There is a module called pylearn2.models.rbm, but it is
    somewhat redundant with pylearn2.models.dbm since an RBM
    is just a single layer DBM. In the future, maybe the RBM
    class should be rewritten as a single layer subclass of
    the DBM.

    For now, you can train an RBM model using either the RBM
    class or the DBM class. This directory contains an example
    of how to do the latter. Just run

    python train_dbm.py

    This sets hyperparameters(hyper_params) and then runs the 
    experiment defined in rbm.yaml with those hyper_parameter values.  

    See rbm.yaml for further comments.

    After you have trained this, you should be able to run some of
    the scripts in pylearn2/scripts/dbm on it. For example, the
    show_negative_chains.py script will let you see the last state
    of the fantasy particles used for the PCD training.

    Some of the more generic scripts in pylearn2/scripts also apply.
    For example, show_weights.py will let you see the learned filters.

