Same delta and beta for all neuron of one layer

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
1

{moderator edit - quiz answers removed}

hi
is not true that we compute mean and variance over the input of one layer z for all neuron and then
add delta and beta to reach final z and this delta and beta are same for all neuron of this special layer?
thanks

2

You might want to watch those lectures again. The point is that the statistics are computed “per neuron” and across the “samples” dimension, right? So those quantities are not one global value, but vectors.

1 Like