The output of a convolutional layer $($internal activation$)$ consists in $256$ feature maps of size $6464,$ this is fed to a $11$ convolutional layer with $128$ filters. The number$-$of$-$feature$-$maps$)$:

$646264$

$6464256$

$6464128$

$128128256$

Moving to the next question prevents changes to this answer.