简单神经网络的搭建结构

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
import matplotlib
matplotlib.use('TkAgg')#terimal

def add_layer(inputs, in_size, out_size,activation_function=None):
    #add one more hidden layer and return the output of this layer
    Weights = tf.Variable(tf.random_normal([in_size,out_size]))
    
    biases = tf.Variable(tf.zeros([1,out_size])+0.1)
    
    Wx_plus_b = tf.matmul(inputs,Weights) + biases
    
    if activation_function is None:
        outputs = Wx_plus_b
    else:
        outputs = activation_function(Wx_plus_b)
    return outputs

#Make up some real data
x_data = np.linspace(-1,1,300)[:,np.newaxis]
noise = np.random.normal(0,0.05,x_data.shape)
y_data = np.square(x_data)-0.5 + noise


#define placeholder for inputs to network
xs = tf.placeholder(tf.float32,[None,1])
ys = tf.placeholder(tf.float32,[None,1])

#add hidden layr
L1 = add_layer(xs,1,10,activation_function=tf.nn.relu)

#add output layer
prediction = add_layer(L1,10,1,activation_function=None)

#opt,loss
loss = tf.reduce_mean(tf.reduce_sum(tf.square(ys - prediction),reduction_indices=[1]))
train_step = tf.train.GradientDescentOptimizer(0.1).minimize(loss)
#init before train
init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init)

#train

for i in range(1000):
    sess.run(train_step,feed_dict={xs: x_data,ys: y_data})
    if i%50==0:
        prediction_value = sess.run(prediction, feed_dict={xs: x_data})
        
        
        print(sess.run(loss,feed_dict={xs: x_data,ys: y_data}))

结果
0.65913
0.013715
0.00760787
0.00660618
0.00610456
0.00574773
0.00545486
0.00519336
0.00497396
0.00479176
0.00464916
0.00451997
0.00440686
0.00430621
0.00421834
0.00415592
0.00409589
0.00404482
0.00397444
0.00390226
ensorflow 中在运行时动态设置某个变量的值，先使用placeholder占位。运行时动态给占位符“喂”数据。

用tf.placeholder占位

import tensorflow as tf

a = tf.placeholder(tf.float32, name = "input_1")
b = tf.placeholder(tf.float32, name = "input_2")
output = tf.multiply(a, b, name = "mul_out")

input_dict = {a : 7.0, b : 10.0}

with tf.Session() as sess:
    print(sess.run(output, feed_dict = input_dict)) #feed_dict是一个字典结构