MATLAB: Neural network training Maximum mu reached

neural network

I am trying to train a neural network. Within first iteration itself it gives maximum mu reached. I have been reading at several places, that It is a sign of convergence. But the model generates a very high error of the order 10^4.

Best Answer

Hello, Matlab selects a default Mu value of 0.001, but you can change it using the command: Net.trainParam.Mu = 0.0001; With this command you can give the value you want before training neural red, if you are using nftool, you can go to the last session called "Save Result" and generate a simple script, and before the line train (net, x , T) change the value of MU, if the error persists, then you should try changing the number of neurons in the hidden layer, or the transfer function of your layers

Related Solutions

MATLAB: How to train a feedfordward neural network with error weights

 Use the command window
 help mse
 doc mse

Hope this helps.

Thank you for formally accepting my answer

Greg

MATLAB: Does train function return neural network with the weights of last epoch or the weights of the model with the smallest loss of all epochs

You can refer to Early Stopping & Improve Shallow Neural Network Generalization and Avoid Overfitting documentation.

According to it "When the validation error increases for a specified number of iterations (net.trainParam.max_fail), the training is stopped, and the weights and biases at the minimum of the validation error are returned."

For more information on the "net.trainParam.max_fail" parameter, enter the following in command window:

help nnparam.max_fail

Best Answer

Related Solutions

MATLAB: How to train a feedfordward neural network with error weights

MATLAB: Does train function return neural network with the weights of last epoch or the weights of the model with the smallest loss of all epochs

Related Question