I cant figure out my feed forward NN’s(neural network) proper dimensions
I am training a neural network with my program for feed forward NN’s and cant figure the amount of hidden layer and nodes The problem: I am training my NN to learn how to add 2 5 long binary lists to get 1, 5 long list (notice I have made sure no two numbers are longer in binary form than 5 digits), everything is good besides the ammount of hidden layers and nodes for them, please help I can provide the code if needed
I've never really built my own NN from scratch. Summarizing: https://stats.stackexchange.com/questions/181/how-to-choose-the-number-of-hidden-layers-and-nodes-in-a-feedforward-neural-netw Layers: There is no practical purpose of having more than two hidden layers, and one hidden layer is sufficient for many purpose. Nodes: Nh = Ns / (a * (Ni + No)) where Nh = number of hidden nodes Ns = number of training data Ni = number of input neurons No = number of output neurons and the scaling factor a was recommended to be between 2 to 10.
im a bit confused on Ns, for my problem it would be over 372827 neurons well, asuming Ns is the ammount of traning data