search

RichardBJ / Deep-Channel

To follow
MIT License
43 stars 19 forks source link

MaxPool size 1 #3

Closed khornlund closed 3 years ago

khornlund commented 4 years ago

Hi could you explain the purpose of the MaxPooling1D layer with size 1? Is this redundant?

numancelik34 commented 4 years ago

Hi @khornlund I think the pool size equals to 1, because both the size of timestep, and input features are 1 in the input shape of 1DCNN, and to apply TimeDistributed.Flatten(), MaxPooling layer is embedded before that. Let me know please if I am not correct!

RichardBJ commented 3 years ago

This sounds resolved so closing. Thanks