《深度学习之Tensorflow：入门、原理与进阶实战》——线性回归

import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' # 屏蔽这个警告，烦

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

# tfe = tf.contrib.eager # 开启Eager Execution
# tf.enable_eager_execution()
# # 在tf 2中，已经用eager来替代了session，使代码更加简洁
# data = tf.constant([1,2])
# print(data.numpy()) # 打印常数格式

plotdata = {"batchsize":[], "loss":[]}
def moving_average(a, w=10):
    if len(a) < w:
        return a[:]
    return [val if idx < w else sum(a[(idx-w):idx])/w for idx, val in enumerate(a)]

# 线性回归 y=2x，加入噪声
train_X = np.linspace(-1, 1, 100)
train_Y = 2 * train_X + np.random.randn(*train_X.shape) * 0.3
# 显示模拟数据点
plt.plot(train_X, train_Y, 'ro', label='Original data')
plt.legend()
# plt.show()

# 创建模型
# 占位符
X = tf.placeholder("float")
Y = tf.placeholder("float")
# 模型参数
W = tf.Variable(tf.random_normal([1], name="weight"))
b = tf.Variable(tf.zeros([1]), name="bias")
# 前向结构
z = tf.multiply(X, W) + b
# 反向优化
cost = tf.reduce_mean(tf.square(Y - z)) # 生成值与真实值的平方差
learning_rate = 0.01 # 学习率，代表调整参数的速度
optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost) # 梯度下降

# 初始化所有变量
init = tf.global_variables_initializer()
# 定义参数
training_epochs = 20 # 迭代次数设置为20
display_step = 2

# 启动session
with tf.Session() as sess:
    sess.run(init)
    plotdata = {"batchsize":[], "loss":[]} #存放批次和损失值
    for epoch in range(training_epochs):
        # 向模型输入数据
        for (x, y) in zip(train_X, train_Y):
            sess.run(optimizer, feed_dict={X:x, Y:y})

        # 显示训练中的详细信息
        if epoch % display_step == 0:
            loss = sess.run(cost, feed_dict={X:train_X, Y:train_Y})
            print("Epoch:", epoch+1, "cost=", loss, "W=", sess.run(W), "b=", sess.run(b))
            if not (loss == "NA"):
                plotdata["batchsize"].append(epoch)
                plotdata["loss"].append(loss)
        
    print("Finished!")
    print("cost=", sess.run(cost, feed_dict={X:train_X, Y:train_Y}), "W=", sess.run(W), "b=", sess.run(b))

    # 图形显示
    plt.plot(train_X, train_Y, 'ro', label='Original data')
    plt.plot(train_X, sess.run(W) * train_X + sess.run(b), label='Fitted line')
    plt.legend()
    plt.show()

    plotdata["avgloss"] = moving_average(plotdata["loss"])
    plt.figure(1)
    plt.subplot(211)
    plt.plot(plotdata["batchsize"], plotdata["avgloss"], 'b--')
    plt.xlabel('Minibatch number')
    plt.ylabel('Loss')
    plt.title('Minibatch run vs. Training loss')

    plt.show()

    print("x=0.2, z=", sess.run(z, feed_dict={X: 0.2}))