General purpose fully-connected MLP in theano
General purpose fully-connected MLP in theano
The functions required to create the MLP are provided in nnetFunc.py. One example of training using cross validation for a multi-output network is given in trainNNetXval.py. This last file can be modified to change the parameters of the model (number of hidden units, hidden layers, regularisation, learning rate,…). After learning, the MLP can be applied to new data as shown in the applyNNet.py.
All the functions contained in this repository were developed in the Multiscale Cardiovascular Engineering Group at University College London (MUSE-UCL) by Carlos Ledezma.
This work is protected by a Creative Commons Attribution-ShareAlike 4.0 International license (https://creativecommons.org/licenses/by-sa/4.0/)