TCN Receptive Field

Problem

Compute the receptive field of a stack of dilated causal convolutions: 1 + (kernel_size - 1) * sum(dilations).

Each dilation multiplies how far the layer looks back, so the field grows quickly with depth.

Hint

With no layers the field is just the current position.

case 1

in: [3,[1,2,4]]

expected: 15

case 2

in: [2,[1,1,1]]

expected: 4

case 3

in: [3,[]]

expected: 1

case 4

in: [5,[1,2]]

expected: 13

Starter codePython

def tcn_receptive_field(kernel_size, dilations):
    # Your code here
    pass

4 test cases · run your code against them in the editor