Gradient Descent Machine Learning

preview_player
Показать описание

Рекомендации по теме
Комментарии
Автор

Learning rate is a hyperparameter(values choose by humens)... We are not finding learning rate via gradient decent anytime. Its used to find the minimum weight of the model

pavithradevim
Автор

choosing a convex function will not give you a correct picture of gradient decent, try for non-convex function, as for gradient decent it will be hard to find minima their

ashish
Автор

Empirical risk or average loss, not loss

marcelgunadi