Tensorflow神经网络框架（第二课 2-4Tensorflow简单实例 线性回归 梯度下降法）

2-4Tensorflow简单实例Last Checkpoint: 上星期五15:41(unsaved changes) Logout

Python 3 

Trusted

• File
• Edit
• View
• Insert
• Cell
• Kernel
• Help

Run

CodeMarkdownRaw NBConvertHeading-

In [1]:

import tensorflow as tf

import numpy as np

In [2]:

#使用numpy生成100个随机点

x_data = np.random.rand(100)

y_data = x_data*0.1 + 0.2 # 构造一条直线 斜率：0.1 

​

#构造一个线性模型

b = tf.Variable(0.)

k = tf.Variable(0.)

y = k*x_data + b

​

#二次代价函数

loss = tf.reduce_mean(tf.square(y_data-y))

#定义一个梯度下降法来进行训练的优化器

optimizer = tf.train.GradientDescentOptimizer(0.2) #学习率：0.2

#最小化代价函数

train = optimizer.minimize(loss)

​

init = tf.global_variables_initializer()

​

with tf.Session() as sess:

    sess.run(init)

    for step in range(201): #迭代201次

        sess.run(train)

        if step%20 == 0:#每迭代20次 打印出来

            print(step,sess.run([k,b]))

0 [0.048383515, 0.09845313]
20 [0.09913179, 0.20042612]
40 [0.099467784, 0.20026132]
60 [0.09967373, 0.20016019]
80 [0.09979998, 0.20009822]
100 [0.09987738, 0.2000602]
120 [0.09992483, 0.20003691]
140 [0.09995392, 0.20002262]
160 [0.099971764, 0.20001386]
180 [0.09998268, 0.20000851]
200 [0.09998938, 0.20000522]