Tensorflow normal

 

import numpy as np
import tensorflow as tf

def main():
    testNormal()
    testPlaceholder()
    testPlaceholder1()
    test3()
    testGet_tensor_by_name()

def testNormal():
    #tf.random_uniform((4, 4), minval=low,maxval=high,dtype=tf.float32)))返回4*4的矩阵，产生于low和high之间，产生的值是均匀分布的。
    #sq = tf.square(tf.constant([2, 4, 6, 8]))
    sq1 = tf.random_uniform((2, 2), -1, 1)
    init = tf.global_variables_initializer()
    
    sess = tf.Session()
    sess.run(init)
    #print(sess.run(sq)) # [ 4 16 36 64]
    print(sess.run(sq1))
    sess.close()
    
def testPlaceholder1():
    input1 = tf.placeholder(tf.float32)
    input2 = tf.placeholder(tf.float32)
    output = tf.multiply(input1, input2)

    with tf.Session() as sess:
        print(sess.run([output], feed_dict={input1:7., input2:2.}))
        print(sess.run([output], feed_dict={input1:[7.], input2:[2.]}))
    
#A Simple TensorFlow example
#用Tensorflow计算a=(b+c)∗(c+2)
def testPlaceholder():
    #1. 定义数据：
    # 首先，创建一个TensorFlow常量=>2
    const = tf.constant(2.0, name='const')
    # 创建TensorFlow变量b和c
    #b = tf.Variable(2.0, name='b')
    # 创建placeholder
    # TensorFlow 还提供了 feed机制, 该机制 可以临时替代图中的任意操作中的 tensor 
    #可以对图中任何操作提交补丁, 直接插入一个 tensor.
    b = tf.placeholder(tf.float32, [None, 1], name='b')
    c = tf.Variable(1.0, dtype=tf.float32, name='c')

    #2. 定义运算(也称TensorFlow operation)：
    d = tf.add(b, c, name='d')
    e = tf.add(c, const, name='e')
    a = tf.multiply(d, e, name='a')
    
    #TensorFlow中所有的变量必须经过初始化才能使用，初始化方式分两步：    
    #<1. 定义初始化operation
    init_op = tf.global_variables_initializer()

    # session
    with tf.Session() as sess:
        #<2. 运行init operation
        sess.run(init_op)
        # 计算
        #a_out = sess.run(a)
        a_out = sess.run(a, feed_dict={b: np.arange(0, 10)[:, np.newaxis]})
        print("Variable a is {}".format(a_out))


def test3():
    a = tf.add(2, 5)     #a=7
    b = tf.multiply(a, 3)    #b=3*7=21
    with tf.Session() as sess:
        print(sess.run(b))  # 21 
        replace_dict = {a:15}   #用15代替b算式中的a
        # TF Session.run注入参数简单例子, feed_dict参数的作用是替换图中的某个tensor的值。
        print(sess.run(b, feed_dict = replace_dict)) #45

def testGet_tensor_by_name():
    c = tf.placeholder(tf.float32, [2, 2], name='namec')  
    d = tf.constant([[1.0, 1.0], [0.0, 1.0]])
    e = tf.matmul(c, d, name='example')

    with tf.Session() as sess:
        result = sess.run(e, feed_dict={c : [[1, 1], [1, 1]]})
        print(result) # [[1. 2.], [1. 2.]]
        print (e.name)  #example:0
        testc = tf.get_default_graph().get_tensor_by_name("namec:0")
        print(testc)          
        print(sess.run(e, feed_dict={testc:[[1, 0], [0, 1]]})) #[[1. 1.], [0. 1.]]        

if __name__ == '__main__':
    main()

[[-0.22921348 0.68460655]
[ 0.21008825 -0.75417924]]
Variable a is [[ 3.]
[ 6.]
[ 9.]
[12.]
[15.]
[18.]
[21.]
[24.]
[27.]
[30.]]
[14.0]
[array([14.], dtype=float32)]
21
45
[[1. 2.]
[1. 2.]]
example:0
Tensor("namec:0", shape=(2, 2), dtype=float32)
[[1. 1.]
[0. 1.]]