In neural networks, You'd want the cost function to reach the global minimum but not slowly, that is why you could sometimes use Adam algorithm as an optimizer to help you adjust the learning rate.
2 weeks ago (edited)
| 0
GD with momentum can go through saddle point and escape from local minimum,sometimes not 😂
1 week ago | 0
Hire Ready
What does the learning rate control in gradient descent?
2 weeks ago | [YT] | 4