Neural network model

Observer

When training the neural network, if you would like to monitor the progress you can do so with an observer. To do so, attach it before you train with something like

def observer(model, vector, loss):
    """This function can extract data during optimization.
    Full access is provided to the model, the vector of parameters, and
    the current value of the loss function.
    """
    pass

calc.model.observer = myobserver

Note that the observer must take exactly the three arguments specified above.

Your function “observer” will be called at each call to the loss function. For example, you can use this to print out values of specific parameter functions.