Function calling from a class

in many machine learning codes, i often see the training network class defined with its init function and forward functions. However, when initializing it, we don't specifically call the forward function(eg. x= network() and then x.forward()), yet the forward function gets implemented. so what is really going on under the hood? is it only in machine learning definitions or there is a general concept that can be applied, like calling a certain function without explicitly . calling the function of the specific class?