Softmax of Logits

Problem

Compute the softmax of a list of logits: exp(z_i - max(z)) / sum(exp(z_j - max(z))).

Return the probability list; empty input returns [].

Hint

Subtract the maximum logit first to avoid overflow, then normalize.

case 1

in: [[1,1,1]]

expected: [0.3333333333333333,0.3333333333333333,0.3333333333333333]

case 2

in: [[0,1]]

expected: [0.2689414213699951,0.7310585786300049]

case 3

in: [[1000,1000]]

expected: [0.5,0.5]

case 4

in: [[]]

expected: []

Starter codePython

import math
def softmax(logits):
    # Your code here
    pass

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