NumPy 배열에 함수를 매핑하는 가장 효율적인 방법은 무엇입니까?

import numpy as np

x = np.array([1, 2, 3, 4, 5])
squares = np.square(x)  # Fast and straightforward

import numpy as np

def my_function(x):
    # Define your custom function

x = np.array([1, 2, 3, 4, 5])
squares = np.array([my_function(xi) for xi in x])  # Reasonably efficient

import numpy as np

def my_function(x):
    # Define your custom function

x = np.array([1, 2, 3, 4, 5])
squares = np.array(list(map(my_function, x)))  # Slightly less efficient

import numpy as np

def my_function(x):
    # Define your custom function
    return iter([my_function(xi) for xi in x])  # Yields values as an iterator

x = np.array([1, 2, 3, 4, 5])
squares = np.fromiter(my_function(x), x.dtype)  # Less efficient, but works with iterators

import numpy as np

def my_function(x):
    # Define your custom function

x = np.array([1, 2, 3, 4, 5])
my_vectorized_function = np.vectorize(my_function)
squares = my_vectorized_function(x)  # Most efficient, but may not always be possible