numpy 三种矩阵乘法multiply, matmul和 dot

版权声明：本文为博主CSDN Rosefun96原创文章。 https://blog.csdn.net/rosefun96/article/details/87894492

1 简介

np.multiply是对应元素相乘，np.matmul和np.dot都是矩阵相乘运算，之间等价。

y_pred = [[0.38574776, 0.08795848, 0.83927506],
          [0.21592768, 0.44453627, 0.10463644],
          [0.8793516,  0.65118235, 0.5184219 ]]
y_true = [[1., 0., 0.],
          [0., 0., 0.],
          [1., 1., 1.]]

print('multiply \n',np.multiply(y_pred, y_true))
print('matmul \n',np.matmul(y_pred, y_true))
print('dot \n',np.dot(y_pred, y_true))

结果：

multiply 
 [[0.38574776 0.         0.        ]
 [0.         0.         0.        ]
 [0.8793516  0.65118235 0.5184219 ]]
matmul 
 [[1.22502282 0.83927506 0.83927506]
 [0.32056412 0.10463644 0.10463644]
 [1.3977735  0.5184219  0.5184219 ]]
dot 
 [[1.22502282 0.83927506 0.83927506]
 [0.32056412 0.10463644 0.10463644]
 [1.3977735  0.5184219  0.5184219 ]]