numpy库中一维矩阵的一些坑点

import numpy as np

w = np.array(range(15)).reshape(3,5)
b = np.array(range(5))
c = b.reshape(1,b.shape[0])
print w.shape,b.shape, c.shape
#(3, 5) (5,) (1, 5)
a = w -b
d = w -c
print a, a.shape
"""
[[ 0  0  0  0  0]
 [ 5  5  5  5  5]
 [10 10 10 10 10]] (3, 5)
 """
print d, d.shape
"""
[[ 0  0  0  0  0]
 [ 5  5  5  5  5]
 [10 10 10 10 10]] (3, 5)
 """

#b += c
#ValueError: non-broadcastable output operand with shape (5,) doesn't match the broadcast shape (1,5)
c += b
print c, c.shape
#[[0 2 4 6 8]] (1, 5)




可以看见b（5，）可以向c（1，5）或其他（n，5）传播，这个传播是单向的这个也有点意思 https://stackoverflow.com/questions/32109319/how-to-implement-the-relu-function-in-numpy