On return, all gradients are store in the appropriate SystemData.d fields
(see system_data).
FUNCTION
This function calculates the gradient of the MSE with respect to
all the trainable parameters in a net. First, it calculates the error
at the output layer. Then, the error is calculated for each unit by
back-propagation, using the grad_unit function.
NOTE
This function is used by the function grad_net_mse.
The code is an adaptation of the back-propagation function
bp_adapt_net.
This document was generated using api2html on Thu Mar 5 09:00:00 MET DST 1998