iCelNNWeights Struct Reference

Structure which can hold the weightings of all links between the nodes of an iPcNeuralNet property class. More...

#include <propclass/neuralnet.h>

Inheritance diagram for iCelNNWeights:

List of all members.

Public Member Functions
virtual csPtr< iCelNNWeights >	Clone () const =0
	Returns a new structure which is a copy of this one.
virtual csArray< csArray < csArray< float > > > &	Data ()=0
	Access the weighting data (non-const). Use with care.
virtual const csArray< csArray < csArray< float > > > &	Data () const =0
	Access the weighting data.

---

Detailed Description

Structure which can hold the weightings of all links between the nodes of an iPcNeuralNet property class.

This is not really intended to be used by user code, only by neural network training code like celPcEvolve.

Data()[i][j][k] is the weighting from node k to node j at layer i.

Warning:

You must only modify the contents of the innermost arrays. Meaning, you must not replace any outer array, for example Data()[i] = myArray to replace an entire layer. If you do, you will cause memory allocation problems. You must instead copy each individual float value separately.

Definition at line 49 of file neuralnet.h.

---

Member Function Documentation

virtual csPtr<iCelNNWeights> iCelNNWeights::Clone

(

)

const [pure virtual]

Returns a new structure which is a copy of this one.

virtual csArray< csArray< csArray<float> > >& iCelNNWeights::Data

(

)

[pure virtual]

Access the weighting data (non-const). Use with care.

virtual const csArray< csArray< csArray<float> > >& iCelNNWeights::Data

(

)

const [pure virtual]

Access the weighting data.

---

The documentation for this struct was generated from the following file:

• propclass/neuralnet.h