C5W3A2-Reasoning Behind the Network Architecture?

Course Q&A Deep Learning Specialization Sequence Models
1

Hi everyone,

I hope you all having fun playing with the Trigger Word Detection lab.
Speaking of which, I am wondering why the network is designed the way it is shown in Figure 3 of the second assignment lab in C5W3. Specifically,
1- Why are there two Dropouts used after the second GRU?
2- Why is there Dropout-BatchNorm after the first GRU, but Droupout-BatchNorm-Droupout after the second GRU? While is there BatchNorm-Dropout after the Conv1D?

I appreciate your advice.

Hossein

2

I suspect that’s what they arrived at as giving suitable performance without requiring too much computational horsepower for a lab assignment.

1 Like