In a Neural Network problem how do we decide the size/dimensions of the parameter Weight (W) and Bias (b)?
The number of units in the hidden layers is determined by experimentation.
There is always one bias weight per unit.
In a Neural Network problem how do we decide the size/dimensions of the parameter Weight (W) and Bias (b)?
The number of units in the hidden layers is determined by experimentation.
There is always one bias weight per unit.