padの使い方 (Python-numpy)

import numpy as np

a = np.array([1,2,3,])

b = np.pad(a,[2,3]) # 配列の前に0を2個、後ろに0を3個加える
print(b)

    → array([0, 0, 1, 2, 3, 0, 0, 0])

b = np.pad(a,[2,3] , constant_values=(5,6)) # 配列の前に5を2個、後ろに6を3個加える
print(b)

    → array([5, 5, 1, 2, 3, 6, 6, 6])

b = np.pad(a,[2,3],'edge')　#端と同じ数値を加える
print(b)

    → array([1, 1, 1, 2, 3, 3, 3, 3])

a = np.array([[1,2,3],[4,5,6]])

b = np.pad(a , (1,2)) #配列の上と前に1つ、下と左に2つ0を加える
print(b)

    → array([[0, 0, 0, 0, 0, 0],
                [0, 1, 2, 3, 0, 0],
                [0, 4, 5, 6, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]])

b = np.pad(a , [(0,1),(2,3)]) #配列の上に0個、下に1つ、前(左)に2つ、後ろ(右)に3つ加える。
print(b)

    →array([[0, 0, 1, 2, 3, 0, 0, 0],
                [0, 0, 4, 5, 6, 0, 0, 0],
                [0, 0, 0, 0, 0, 0, 0, 0]])

a=np.array([[[1,2],[3,4]],[[5,6],[7,8]]])
print(a)

    → array([[[1, 2],
                [3, 4]],

                [[5, 6],
                [7, 8]]])

b = np.pad(a , [(1,1),(2,3),(1,2)]) #配列の前後が(1,1)、配列の中の上下が(2,3)、配列の中の前後が(1,2)
print(b)

    → array([[[0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],

                [[0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 1, 2, 0, 0],
                [0, 3, 4, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 5, 6, 0, 0],
                [0, 7, 8, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0]]])

a=np.array([[[1,2],[3,4]],[[5,6],[7,8]]]
a=a.reshape(2,2,2,1)
print(a)

    → array([[[[1],
                [2]],

                [[3],
                [4]]],


                [[[5],
                [6]],

                [[7],
                [8]]]])

b = np.pad(a , [(1,0),(0,1),(1,1),(2,3)])
print(b)

    → array([[[[0, 0, 0, 0, 0, 0],　#3つの配列とも、(1,0)の1の部分,
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]]],


                [[[0, 0, 0, 0, 0, 0],　#上下が(1,1)、前後が(2,3)
                [0, 0, 1, 0, 0, 0],
                [0, 0, 2, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0, 0],
                [0, 0, 3, 0, 0, 0],
                [0, 0, 4, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0, 0],　#(0,1)の1の部分
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]]],


                [[[0, 0, 0, 0, 0, 0],
                [0, 0, 5, 0, 0, 0],
                [0, 0, 6, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0, 0],
                [0, 0, 7, 0, 0, 0],
                [0, 0, 8, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]],

                [[0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0],
                [0, 0, 0, 0, 0, 0]]]])