TensorFlow训练模型

import tensorflow as tf
import numpy as np
# 创建数据
x_data = np.random.rand(100).astype(np.float32)
y_data = x_data * 0.1 + 0.3
# 搭建模型
Weight = tf.Variable(tf.random_uniform([1], -1.0, 1.0))
# Variable 用来创建变量的参数
biases = tf.Variable(tf.zeros([1]))
# 学着把Weight 变成0.1  biases 变成0.3
y = Weight * x_data + biases

loss = tf.reduce_mean(tf.square(y - y_data))  # 计算y 和 y_data 的误差
optimizer = tf.train.GradientDescentOptimizer(0.5)  # 0.5自己进行定义
train = optimizer.minimize(loss)  # 梯度下降法进行训练

init = tf.global_variables_initializer()  # 建立了神经网络结构，必须初始化所有之前定义的Variable

sess = tf.Session()  # 逐步提升预测的准确性
sess.run(init)  # 非常重要

for step in range(201):
    sess.run(train)
    if step % 20 == 0:
        print(step, sess.run(Weight), sess.run(biases))

import tensorflow as tf
import numpy as np
# 创建数据
x_data = np.random.rand(100).astype(np.float32)
y_data = x_data * 0.1 + 0.3
# 搭建模型
Weight = tf.Variable(tf.random_uniform([1], -1.0, 1.0))
# Variable 用来创建变量的参数
biases = tf.Variable(tf.zeros([1]))
# 学着把Weight 变成0.1  biases 变成0.3
y = Weight * x_data + biases

loss = tf.reduce_mean(tf.square(y - y_data))  # 计算y 和 y_data 的误差
optimizer = tf.train.GradientDescentOptimizer(0.5)  # 0.5自己进行定义
train = optimizer.minimize(loss)  # 梯度下降法进行训练

init = tf.global_variables_initializer()  # 建立了神经网络结构，必须初始化所有之前定义的Variable

# sess = tf.Session()  # 逐步提升预测的准确性
# sess.run(init)  # 非常重要
with tf.Session() as sess:  # 用上下文进行管理  不用关闭
    sess.run(init)
    for step in range(201):
        sess.run(train)
        if step % 20 == 0:
            print(step, sess.run(Weight), sess.run(biases))

输出结果

0 [-0.24398923] [0.653689]

20 [0.00985567] [0.34723642]

40 [0.08043306] [0.31025326]

60 [0.09575275] [0.30222562]

80 [0.0990781] [0.3004831]

100 [0.09979989] [0.3001049]

120 [0.09995657] [0.30002278]

140 [0.09999057] [0.30000496]

160 [0.09999795] [0.30000108]

180 [0.09999954] [0.30000025]

200 [0.0999999] [0.30000007]