Package crino :: Module module :: Class Sequential

Class Sequential

Known Subclasses:

network.MultiLayerPerceptron

A Sequential module computes its outputs sequentially, i.e. each outputs of its submodules is linked to the inputs of the following module. The number of submodules in the sequence can be chosen arbitrarily, but the inputs and outputs sizes must be the same throughout the sequence.

Instance Methods

[hide private]

__init__(self, mods=[], nInputs=None)
Constructs a new Sequential container. source code

prepareGeometry(self)
Sets the same inputs and outputs size for all submodules, and prepare their internal geometry.

source code

prepareParams(self)
Initializes the params of the submodules.

source code

prepareOutput(self)
Computes sequentially the symbolic outputs of the module.

source code

Inherited from Container: add

Inherited from Module: criterionFunction, forward, forwardFunction, holdFunction, linkInputs, linkModule, prepare, prepareBackup, restoreFunction, save, trainFunction

Instance Variables

[hide private]

Inherited from Container: modules

Inherited from Module: backupParams, inputs, nInputs, nOutputs, outputs, params, prepared

Method Details

[hide private]

init(self, mods=`[]`, nInputs=None)
(Constructor)

source code

Constructs a new Sequential container.

Parameters:

mods (list) - A list of submodules to add to the sequence. They will be linked in the same order as provided.
nInputs (int) - The inputs size of the sequence (and of all the submodules).

Overrides: Module.__init__