tf.matmul(),tf.multiply(),np.dot

import tensorflow as tf

import numpy as np

x=tf.constant([[1,2,3],[1,2,3],[1,2,3]])

y=tf.constant([[2,1,1],[2,1,1],[2,1,1]])

x1=([[1,2,3],[1,2,3],[1,2,3]])

y1=([[2,1,1],[2,1,1],[2,1,1]])

z=tf.multiply(x,y)

z1=tf.matmul(x,y)

z2 = np.dot(x1,y1)

print('dot\n',z2)

z3 = np.multiply(x1,y1)

print('np.multiply\n',z3)

with tf.Session() as sess:

print('tf.multiply\n',sess.run(z))

print('tf.matmul\n',sess.run(z1))

print(sess.run(np.dot(x,y)))

'''

之前使用sess.run(np.dot(x,y))输出结果为点乘结果，直接计算np.dot(x1,y1)结果为矩阵乘法.

tf.matmul()  为矩阵乘法

tf.multiply() 为矩阵点乘

np.dot() 为矩阵乘法

np.multiply() 为矩阵点乘

=============== 输出结果 ===============

np.dot

[[12  6  6]

[12  6  6]

[12  6  6]]

np.multiply

[[2 2 3]

[2 2 3]

[2 2 3]]

tf.multiply

[[2 2 3]

[2 2 3]

[2 2 3]]

tf.matmul

[[12  6  6]

[12  6  6]

[12  6  6]]

'''