gradient-optimizers 0.0.1

Python package for wrapping gradient optimizers for models in Theano

Two classes GradientModel, and GradientHFModel, for optimizing gradient
based models (specifically built with indexed parameters in mind (e.g.
for language models))

GradientModel
-------------

A gradient model for updating your model with
hessian free, adagrad, or linear decay updates.

You will need to define the following attributes,
and fill them as appropriate:

self.params = []
self.indexed_params = set()

self._l2_regularization = True / False

# if L2 is true store this parameter:
self._l2_regularization_parameter = theano.shared(np.float64(l2_regularization).astype(REAL), name='l2_regularization_parameter')

Upon initialization you must run:

self._select_update_mechanism(update_method_name)

# then to compile this mechanism:
self.create_update_fun()


The update methods expect the input to be of the form:

ivector <indices/>, ivector <labels/>

If this is not the case you can modify them as appropriate.

GradientHFModel
---------------

Implements an symbolic one step of hessian-free [1]
optimization that approximates the curvature,
requires a _compute_cost method that takes an example
as input or a _compute_cost_gradients that returns
gradients for each example provided.

Model should have a params property containing symbolic
theano variables.

[1] James Martens, ``Deep learning via Hessian-free optimization", ICML 2010

Make sure the following parameters are not tampered with:

self._additional_params

self._num_updates