How to insert dropout layers after activation layers in a pre-trained non-sequential model using functional keras API?

Question

I am working on a modified resnet, and want to insert dropout after activation layers. I have tried this but due to the model not being sequential, it did not work: