numpy[..., None]的理解

numpy数组维度

1	import numpy as np

1 2	arr_1 = np.array([1, 2, 3, 4]) print(arr_1, '\n' , 'shape of arr_1:', arr_1.shape, '， dimension of arr_1:',np.ndim(arr_1))

output:
arr_1 = [1 2 3 4] 
shape of arr_1: (4,) ， dimension of arr_1: 1

1 2	arr_2 = np.array([[1, 2, 3, 4], [5, 6, 7, 8]]) print(arr_2, '\n' , 'shape of arr_2:', arr_2.shape, '， dimension of arr_2:',np.ndim(arr_2))

output:
arr_2 = 
    [[1 2 3 4]
     [5 6 7 8]] 
shape of arr_2: (2, 4) ， dimension of arr_2: 2

1
2

arr_3 = np.array([[[1, 2, 3, 4], [5, 6, 7, 8]], [[1, 2, 3, 4], [5, 6, 7, 8]], [[1, 2, 3, 4], [5, 6, 7, 8]]])
print(arr_3, '\n' , 'shape of arr_3:', arr_3.shape, '， dimension of arr_3:',np.ndim(arr_3))

output:
arr_3 = 
    [[[1 2 3 4]
      [5 6 7 8]]

     [[1 2 3 4]
      [5 6 7 8]]

     [[1 2 3 4]
      [5 6 7 8]]]
shape of arr_3: (3, 2, 4) ， dimension of arr_3: 3

numpy数组切片中[…]的理解

假设 x 是一个数组，np.ndim(x) == 5

x[1,2,...] == x[1,2,:,:,:]
x[...,3] == x[:,:,:,:,3]
x[4,...,5,:] == x[4,:,:,5,:]

numpy数组切片中None的理解

None 的作用就是在相应的位置上增加了一个维度，在这个维度上只有一个元素

假设 x.shape == (a, b)，则

(a, b) ==> [None, :, :] ==> (1, a, b)
(a, b) ==> [:, None, :] ==> (a, 1, b)
(a, b) ==> [:, :, None] ==> (a, b, 1)

1
2
3

import numpy as np
arr = np.array([[1,2,3],[4,5,6]])
print(arr, '\n' , 'shape of arr:', arr.shape, '， dimension of arr:',np.ndim(arr))

output:
arr = 
    [[1 2 3]
     [4 5 6]] 
shape of arr: (2, 3) ， dimension of arr: 2

1 2	None_1 = arr[None, :, :] print(None_1, '\n' , 'shape of None_1:', None_1.shape, '， dimension of None_1:',np.ndim(None_1))

output:
None_1 = 
    [[[1 2 3]
      [4 5 6]]] 
shape of None_1: (1, 2, 3) ， dimension of None_1: 3

1 2	None_2 = arr[:, None, :] print(None_2, '\n' , 'shape of None_2:', None_2.shape, '， dimension of None_2:',np.ndim(None_2))

output:
None_2 = 
    [[[1 2 3]]

     [[4 5 6]]] 
shape of None_2: (2, 1, 3) ， dimension of None_2: 3

1 2	None_3 = arr[:, :, None] print(None_3, '\n' , 'shape of None_3:', None_3.shape, '， dimension of None_3:',np.ndim(None_3))

output:
None_3 = 
    [[[1]
      [2]
      [3]]

     [[4]
      [5]
      [6]]] 
shape of None_3: (2, 3, 1) ， dimension of None_3: 3

numpy[…, None]的理解

1 2	None_3 = arr[..., None] # 等价于 None_3 = arr[:, :, None] print(None_3, '\n' , 'shape of None_3:', None_3.shape, '， dimension of None_3:',np.ndim(None_3))

output:
None_3 = 
    [[[1]
      [2]
      [3]]

     [[4]
      [5]
      [6]]] 
shape of None_3: (2, 3, 1) ， dimension of None_3: 3

1 2	y = np.arange(12).reshape((2,2,3)) print(y, '\n' , 'shape of y:', y.shape, '， dimension of y:',np.ndim(y))

output:
y = 
    [[[ 0  1  2]
      [ 3  4  5]]

     [[ 6  7  8]
      [ 9 10 11]]] 
shape of y: (2, 2, 3) ， dimension of y: 3

1 2	y = y[..., None] print(y, '\n' , 'shape of y:', y.shape, '， dimension of y:',np.ndim(y))

output:
y = 
    [[[[ 0]
       [ 1]
       [ 2]]

      [[ 3]
       [ 4]
       [ 5]]]

     [[[ 6]
       [ 7]
       [ 8]]

      [[ 9]
       [10]
       [11]]]] 
shape of y: (2, 2, 3, 1) ， dimension of y: 4