Hi, say the keep_prob = 0.5 for a hidden layer with 5 hidden units. The chance that all the units are dropped is (0.5)^5 = 0.03125. In 100 iterations, it will occur ~3 times. Is the whole layer shut down then? How does the algorithm deal with it? Thanks!

Question about dropout process

rmwkwok Closed February 20, 2023, 12:15am 2

Topic		Replies	Views
Question about the dropout process Improving Deep Neural Networks: Hyperparameter tun coursera-platform	10	558	February 28, 2023
Inverted dropout Intuition? Improving Deep Neural Networks: Hyperparameter tun coursera-platform	3	694	May 24, 2022
Implementation keep_prob in dropout Improving Deep Neural Networks: Hyperparameter tun coursera-platform	4	731	July 4, 2021
Dropout limitation Improving Deep Neural Networks: Hyperparameter tun week-module-1 , coursera-platform	1	15	February 23, 2025
Regularization Improving Deep Neural Networks: Hyperparameter tun coursera-platform	3	600	July 15, 2023