Quantization Step Size

Problem

Compute the asymmetric quantization step size: (max(values) - min(values)) / (2^bits - 1). A constant list gives 0.0.

The signature is quantization_step_size(values, bits=8).

Hint

The step is the distance between adjacent representable levels.

case 1

in: [[0,1],8]

expected: 0.00392156862745098

case 2

in: [[-1,1],8]

expected: 0.00784313725490196

case 3

in: [[5],4]

expected: 0

case 4

in: [[-2,3],16]

expected: 0.00007629510948348211

Starter codePython

def quantization_step_size(values, bits=8):
    # Your code here
    pass

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