numpy(ndarray)和pytorch(tensor)对应点相乘、矩阵相乘操作对比总结_ndarray和tensor

1. numpy(ndarray)

A = np.array([[1, 2], [-1, 4]])
B = np.array([[2, 0], [3, 4]])
A
array([[ 1,  2],
       [-1,  4]])
B
array([[2, 0],
       [3, 4]])
------------------------------
# 对应点相乘、点乘
A * B
array([[ 2,  0],
       [-3, 16]])
 
np.multiply(A, B)
array([[ 2,  0],
       [-3, 16]])
------------------------------
# 矩阵相乘、点积、内积
np.dot(A, B)
array([[ 8,  8],
       [10, 16]])

2. pytorch (tensor)

import torch
A = torch.randint(0, 6, (2, 2))
B = torch.randint(0, 6, (2, 2))
A
tensor([[2, 1],
        [4, 5]])
B
tensor([[0, 3],
        [3, 4]])
---------------------------------
# 对应点相乘（点乘）
A.mul(B)
tensor([[ 0,  3],
        [12, 20]])
 
# 矩阵相乘、点积、内积
A.mm(B)
tensor([[ 3, 10],
        [15, 32]])