hi,
i am training a NN and i would like to know how i can make sure that the NN output on which i will cut is between [0,1].
I can of course apply a sigmoide transformation on the output value to have it between 0 and 1). but this is pretty inconvenient, since the output neuron value is ~ 1 (but not within [0,1]) .
If i could access the max and min values for the output neuron, i could apply a linear transformation to have its value within [0,1]
Is there a simple way to directly set the output neuron to TNeuron::kSigmoid, instead of kLinear?
The only solution i found so far, was to add before the output layer one hidden layer with one node for which i know that the neuron is a TNeuron::kSigmoid right before.
thanks for your help, nabil.