numpy.tile为3d数组

Question

我必须从两个值（60,80）获得3D数组维数（4,2,1）

import numpy as np
array([[[ 60.],
        [ 60.]],
   [[ 60.],
    [ 60.]],
   [[ 80.],
    [ 80.]],
   [[ 80.],
    [ 80.]]])

我做如下

a = np.array([[[ 60.]],
       [[ 80.]]])
np.tile(a, (2, 2,1))

并获取

array([[[ 60.],
        [ 60.]],
       [[ 80.],
        [ 80.]],
       [[ 60.],
        [ 60.]],
       [[ 80.],
        [ 80.]]])

如何获得预期的数组？

Answer 1

您可以将合并的np.tile与np.repeat：

结合使用

a = np.array([[[ 60.]],
       [[ 80.]]])
np.repeat(np.tile(a, (1,2,1)), 2, axis=0)

#array([[[ 60.],
#        [ 60.]],

#       [[ 60.],
#        [ 60.]],

#       [[ 80.],
#        [ 80.]],

#       [[ 80.],
#        [ 80.]]])

或使用np.repeat + reshape：

np.repeat(a, 4).reshape(4,2,1)
#array([[[ 60.],
#        [ 60.]],

#       [[ 60.],
#        [ 60.]],

#       [[ 80.],
#        [ 80.]],

#       [[ 80.],
#        [ 80.]]])