对numpy维度有时不是太清楚，经常搞错，写几个例子看一遍就懂了
import numpy as np
a = np.arange(6).reshape([3,2])
print(a)
[[0 1]
[2 3]
[4 5]]
print(a.shape)
(3, 2)
(n,)数组与(n,1)数组的互转
reshape进行维度的转换
a = np.array([1,2,3])
print(a.shape)
(3,)
b = a.reshape(-1,1)
print(b.shape)
(3, 1)
把shape中为1的维度去掉
a = np.array([[1],[2],[3]])
print(a.shape)
(3, 1)
b = np.squeeze(a)
print(b.shape)
(3,)
添加新维度
a = np.array([1,2,3])
b = a[np.newaxis,:]
print(b)
[[1 2 3]]
print(b.shape)
(1, 3)
c = a[:,np.newaxis]
print(c)
[[1]
[2]
[3]]
print(c.shape)
(3, 1)
矩阵乘法
(n,)数组求内积
a = np.array([1,2,3])
print(a.dot(a))
14
(n,1)数组求内积
a = np.array([[1],[2],[3]])
print(a.dot(a))
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-8-be6ff70f8c84> in <module>()
1 a = np.array([[1],[2],[3]])
----> 2 print(a.dot(a))
ValueError: shapes (3,1) and (3,1) not aligned: 1 (dim 1) != 3 (dim 0)
numpy的转置
>>> x = np.arange(4).reshape((2,2))
>>> x
array([[0, 1],
[2, 3]])
>>> np.transpose(x)
array([[0, 2],
[1, 3]])
>>> x = np.ones((1, 2, 3))
>>> np.transpose(x, (1, 0, 2)).shape
(2, 1, 3)
numpy数组展平
a = np.arange(24).reshape(2,3,4)
print(a.shape)
(2, 3, 4)
print(a.flatten().shape)
(24,)
print(a.flatten())
[ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23]
np.sum的axis
a = np.array([[1,2,3],[4,5,6]])
# 对所有数字求和
print(np.sum(a))
21
# 对列求和
print(np.sum(a,axis=0))
[5 7 9]
# 对行求和
print(np.sum(a,axis=1))
[ 6 15]
# 对最后一个维度求和
print(np.sum(a,axis=-1))
[ 6 15]
np.sum的keepdims
a = np.array([[1,2,3],[4,5,6]])
print(np.sum(a,axis=1))
[ 6 15]
print(np.sum(a,axis=1,keepdims=True))
[[ 6]
[15]]
>>> x=np.array([[1001, 1002], [3, 4]])
>>> x -= np.max(x)
>>> x
array([[ -1, 0],
[-999, -998]])
>>> x=np.array([[1001, 1002], [3, 4]])
>>> x -= np.max(x, axis=1)
>>> x
array([[ -1, 998],
[-999, 0]])
>>> import numpy as np
>>> x=np.array([[1001, 1002], [3, 4]])
>>> x -= np.max(x, axis=1, keepdims=True)
>>> x
array([[-1, 0],
[-1, 0]])
