In a way this is already how the model is trained. Model makes a prediction, loss function calculates how “wrong” the prediction was, and we update the weights

In a way this is already how the model is trained. Model makes a prediction, loss function calculates how “wrong” the prediction was, and we update the weights of the model to minimize the loss.