标签:
function cnet = adapt_dw(cnet,dW)%更改各层的权值
%ADAPT_DW apply calculated weights and biases gradient
%
% Syntax
%
% cnet = adapt_dw(cnet,dW)
%
% Description
% Input:
% cnet - Convolutional neural network class object
% dW - delta weights and biases as a single vector
% Output:
% cnet - convolutional neural network with updated weights and biases
% Usage:
% Inorder to update weights and biases
%© Sirotenko Mikhail, 2009
%Weight pointer in reverse order (1 means the bias of the last neuron of last layer)
wPtr = 1;
%==========First all F-layers
for k=cnet.numLayers
标签:
来源: https://blog.csdn.net/tigerc0/article/details/82530349