Neural network optimization

Question

I am trying to understand the purpose of a function in keras. ReduceLROnPlateau()

I realized that this feature helps to reduce the learning speed when there is no improvement in checking for loss. But will the network not get out of the local minimum? What if the network remains at a local minimum of about 5 epochs, and this function further reduces the learning speed, and an increase in the learning speed will actually help the network get out of such a local minimum?

In other words, how will he understand if he has reached a local minimum or plateau?

Answer 1

, CS231n, :

. , , , , . , : , , . , , .

, , . , , . , , , .. . , , . , , , .

, , 10 . , " " " " - , ReduceLROnPlateau .