如何沿某个轴用零填充张量（Python）
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/19349410/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
How to pad with zeros a tensor along some axis (Python)
提问by m4linka
I would like to pad a numpy tensor with 0 along the chosen axis.
For instance, I have tensor rwith shape (4,3,2)but I am only interested in padding only the last two axis (that is, pad only the matrix). Is it possible to do it with the one-line python code?
我想沿着所选轴用 0 填充一个 numpy 张量。例如，我有r形状的张量，(4,3,2)但我只对仅填充最后两个轴感兴趣（即仅填充矩阵）。是否可以使用一行 python 代码来实现？
采纳答案by ali_m
You can use np.pad():
您可以使用np.pad()：
a = np.ones((4, 3, 2))
# npad is a tuple of (n_before, n_after) for each dimension
npad = ((0, 0), (1, 2), (2, 1))
b = np.pad(a, pad_width=npad, mode='constant', constant_values=0)
print(b.shape)
# (4, 6, 5)
print(b)
# [[[ 0. 0. 0. 0. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 0. 0. 0.]
# [ 0. 0. 0. 0. 0.]]
# [[ 0. 0. 0. 0. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 0. 0. 0.]
# [ 0. 0. 0. 0. 0.]]
# [[ 0. 0. 0. 0. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 0. 0. 0.]
# [ 0. 0. 0. 0. 0.]]
# [[ 0. 0. 0. 0. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 1. 1. 0.]
# [ 0. 0. 0. 0. 0.]
# [ 0. 0. 0. 0. 0.]]]
回答by cswu
This function would pad at the end of certain axis.
If you wish to pad both side, just modify it.
此功能将在特定轴的末端填充。
如果你想垫两边，只需修改它。
def pad_along_axis(array: np.ndarray, target_length: int, axis: int = 0):
pad_size = target_length - array.shape[axis]
if pad_size <= 0:
return array
npad = [(0, 0)] * array.ndim
npad[axis] = (0, pad_size)
return np.pad(array, pad_width=npad, mode='constant', constant_values=0)
example:
例子：
>>> a = np.identity(5)
>>> b = pad_along_axis(a, 7, axis=1)
>>> print(a, a.shape)
[[1. 0. 0. 0. 0.]
[0. 1. 0. 0. 0.]
[0. 0. 1. 0. 0.]
[0. 0. 0. 1. 0.]
[0. 0. 0. 0. 1.]] (5, 5)
>>> print(b, b.shape)
[[1. 0. 0. 0. 0. 0. 0.]
[0. 1. 0. 0. 0. 0. 0.]
[0. 0. 1. 0. 0. 0. 0.]
[0. 0. 0. 1. 0. 0. 0.]
[0. 0. 0. 0. 1. 0. 0.]] (5, 7)
