Look at the following Keras code and choose the correct answer:

$```$

model $=$ Sequential$()$

model.add$($Dense$(512,$ activation$=$'relu', input$\_$dim$=3))$

model.add$($Dense$(512,$ activation$=$'relu'$))$

model.add$($Dense$(1))$

model.compile$($optimizer$=$'adam', loss$=$'mae', metrics$=[\text{'}$mae$\text{'}])$

$```$

A$.$ The code above is useless because there is an optimizer $($adam$)$ but no activation function at output layer to optimize.

B$.$ The code above is for classification but not necessary binary because the loss function is not 'binary$\_$crossentropy'

C$.$ The code above is for regression because there is no activation function at the output layer

D$.$ The code above is for binary classification because there is one neuron on the output layer