TensorFlow编程训练11——线性模型复习

import tensorflow as tf 
import os

os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'

# 创建变量W和b节点，并设置初始值
W = tf.Variable([.1], dtype=tf.float32)
b = tf.Variable([-.1], dtype=tf.float32)
# 创建x节点，用来输入实验数据
x = tf.placeholder(tf.float32)
# 创建线性模型
l_model = W * x + b

# 创建y节点，用来输入实验中得到的输出数据，用于损失函数的计算
y = tf.placeholder(tf.float32)

# 创建损失模型
loss = tf.reduce_sum(tf.square(l_model - y))

# 创建Session
sess = tf.Session()
writer = tf.summary.FileWriter("logs/", sess.graph)
# init
init = tf.global_variables_initializer()
sess.run(init)
# 打印测试
# print(sess.run(W))
# print(sess.run(b))

# 创建一个梯度下降优化器，学习率为0.001
optimizer = tf.train.GradientDescentOptimizer(0.001)
train = optimizer.minimize(loss)

# 用两个数组保存训练数据
x_train = [1, 2, 3, 6, 8]
y_train = [4.8, 8.5, 10.4, 21.0, 25.3]

# train loop
for i in range(10000):
    sess.run(train, {x: x_train, y: y_train})
   

# 训练后的结果
print('W: %s b: %s loss: %s' % (sess.run(W), sess.run(b), sess.run(loss, {x: x_train, y: y_train})))