| Neural Network Toolbox | |
By-weight-and-bias layer initialization function
Syntax
Description
initwb is a layer initialization function that initializes a layer's weights and biases according to their own initialization functions.
initwb(net,i) takes two arguments,
net - Neural network.
i - Index of a layer.
and returns the network with layer i's weights and biases updated.
Network Use
You can create a standard network that uses initwb by calling newp or newlin.
To prepare a custom network to be initialized with initwb:
net.initFcn to 'initlay'. (This will set net.initParam to the empty matrix [ ] since initlay has no initialization parameters.)
net.layers{i}.initFcn to 'initwb'.
net.inputWeights{i,j}.learnFcn to a weight initialization function. Set each net.layerWeights{i,j}.learnFcn to a weight initialization function. Set each net.biases{i}.learnFcn to a bias initialization function. (Examples of such functions are rands and midpoint.)
To initialize the network, call init.
See newp and newlin for training examples.
Algorithm
Each weight (bias) in layer i is set to new values calculated according to its weight (bias) initialization function.
See Also
| | initnw | initzero | |