Hacker News new | ask | show | jobs
by janalsncm 849 days ago
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.