Conv FLOPs Count

Problem

Count the multiply-accumulate floating point operations of a 2D convolution.

Return 2 * cin * cout * k * k * h_out * w_out (bias excluded), where the factor 2 accounts for each multiply and add.

case 1

in: [3,16,3,28,28]

expected: 677376

case 2

in: [1,1,3,5,5]

expected: 450

case 3

in: [64,128,3,14,14]

expected: 28901376

Starter codePython

def conv_flops(cin, cout, k, h_out, w_out):
    # Your code here
    pass

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