The AI Interview - Master AI/ML Interviews

In machine learning and statistics, the softmax function is a generalization of the logistic function that converts a vector of scores into probabilities. The log-softmax function is the logarithm of the softmax function, and it is often used for numerical stability when computing the softmax of large numbers.

Given a 1D numpy array of scores, implement a Python function to compute the log-softmax of the array.

Examples

Example 1:

Input:

A = np.array([1, 2, 3])
print(log_softmax(A))

Output: array([-2.4076, -1.4076, -0.4076])

Explanation: The log-softmax function is applied to the input array [1, 2, 3]. The output array contains the log-softmax values for each element.

Starter Code

import numpy as np

def log_softmax(scores: list) -> np.ndarray:
	# Your code here
	pass

Implementation of Log Softmax Function

Examples

Starter Code