平均 numpy 数组超过 for 循环？

Question

我在 for 循环的每次迭代中计算一个 numpy 数组。 我如何平均呢？

例如：

for i in range(5):
    array_that_I_calculate = some_calculation(x,y,z)

Answer 1

尝试这个 -

1. array_that_I_calculate每次迭代时将 array_that_I_calculate 为list_of_arrays
2. 循环结束后，取np.average() of list_of_arrays over axis=0

import numpy as np


##### IGNORE #####
#dummy function that returns (2000,1) array

def some_calculation(x=None,y=None,z=None)
    return np.random.random((2000,1))


##### SOLUTION #####

list_of_arrays = []                                  #<-----

for i in range(5):
    array_that_I_calculate = some_calculation(x,y,z)
    list_of_arrays.append(array_that_I_calculate)    #<-----
    
averaged_array = np.average(list_of_arrays, axis=0)  #<-----
print(averaged_array.shape)

(2000,1)