How can I use my own initialization function in the Neural Network Toolbox?

Question

I need to create my own initialization function for my network. I would like to use something other than INITNW and INITWB.

Neeta Dsouza · Accepted Answer

When a neural network is created with NEWFF, it creates an initialization function called INITLAY. When the following code is executed:
 
 
P = [0 1 2 3 4 5 6 7 8 9 10];

T = [0 1 2 3 4 3 2 1 2 3 4];

net = newff(P,T,5);

net.initFcn
The result is:
 
 
 ans =

 initlay
This is the function that assigns the initialization function of each layer of the network to the function:
 
 
net.layers{1}.initfcn

 ans =

 initnw
This means that layer 1 has an initialization function 'initnw'.
To assign a custom function as the initialization function for a particular layer, simply write up the function and assign it to net.layers{1}.fcn:
 
net.layers{1}.initfcn = 'myfun';
That will assign the first layer's initialization function to the custom function. The rest of the layers can also be given initialization functions in the same way.
 
Run INIT on your network to initialize it. This basically calls INITLAY, which then calls each layer's initialization function:
 
 
init(net)

How can I use my own initialization function in the Neural Network Toolbox?

How to Plot data in excel sheet using MATLAB ?

How can I use LSTM networks to solve a time series regression problem?

How to create an attention layer for deep learning networks?

Nvidia NVLink in MATALB using multiple Geforce 2080 ti cards

How do I build TensorFlow Lite for Deep Learning C++ code generation and deployment?

How to train labels for image blocks using neural network?

Modeling and Simulation of an Induction Motor Using Sliding Mode Control in MATLAB

Solar Inverter with BMS in MATLAB Simulink Complete Simulation Project

Simulation and Kinematic Analysis of a 5-DOF RRRRRP Robotic Manipulator Using MATLAB

Get a Free Consultation or a Sample Assignment Review!