Why is the activation function in the convolution layer is omitted?

[MT010104]

Hello @hasanferit ！
When running the IDC framework, we find that the the activation function is considered in the Dense layer but not in the Convolution layer. I wonder why this step is skipped in the Conv layer.

In addition, we also want to know how to initialize the relevance score of regression models!
Looking forward to your reply. Thanks in advance！:）