C4_W4_Ungraded_Lab_2 Residual

Course Q&A Natural Language Processing NLP with Attention Models
1

I’m having a problem figure it out how to solve

Part 1.2 Residual Model

OK, your turn. Write a function ‘MyResidual’, that uses tl.Branch and tl.Add to build a residual layer.
I’ve no idea what to do. I looked at your example, but in yours when using tl.fn you define an input and an output, and the layer is assigned to a variable that is then added to tl.branch.
From documentation, tl.Add takes 2 tensors, but MyResidual has only layer parameter passed to it.
Any help will be appreciated.

def MyResidual(layer):
    return tl.Serial(
        ### START CODE HERE ###
        tl.Add(),
        tl.branch()
        ### END CODE HERE ###
    )
2

You can try this:

    tl.Branch(layer, None),
    tl.Add()

)