Softmax
MediumMachine LearningMathArray
Description
Given an array of numbers logits, return the softmax probabilities (exp of each value divided by the sum of all exps), each rounded to 4 decimal places.
Examples
Input:
logits = [1,2,3]Output:
[0.09,0.2447,0.6652]Explanation:
Exponentiating and normalizing the values into probabilities gives [0.09,0.2447,0.6652].
Input:
logits = [0,0,0]Output:
[0.3333,0.3333,0.3333]Explanation:
Exponentiating and normalizing the values into probabilities gives [0.3333,0.3333,0.3333].
Input:
logits = [2,1]Output:
[0.7311,0.2689]Explanation:
Exponentiating and normalizing the values into probabilities gives [0.7311,0.2689].
Constraints
- •
1 ≤ logits.length ≤ 10³