np.pad数组填充备忘

参考自numpy文档：numpy.pad — NumPy v1.21 Manual

例子：

a = [[1, 2], [3, 4]]

np.pad(a, ((0, 2), (0, 2)), mode = 'constant')

array([[1, 2, 0, 0],
       [3, 4, 0, 0],
       [0, 0, 0, 0],
       [0, 0, 0, 0]])

numpy.pad(array, pad_width, mode='constant', **kwargs)

        pad_width{sequence, array_like, int}

        Number of values padded to the edges of each axis. ((before_1, after_1), …         (before_N, after_N)) unique pad widths for each axis. ((before, after),) yields same         before and after pad for each axis. (pad,) or int is a shortcut for before = after = pad         width for all axes.

我们在例子中的设置的 pad_width 为 ((0, 2), (0, 2))

对应文档中描述((before_1, after_1), ((before_2, after_2))

1就是按行，2就是按列

那么也就是按行填充(0,2)，顶部填充0行，底部填充两行

按列填充(0,2)，左边填充0列，右边填充两列