20210115

今天是写一个线性回归的实例。

但是由于TF1.X和TF2.X的区别，所给的教学视频里的实例已完全不可用(除非禁用掉全部TF2.0特性)，因而转为自行找到了一个实例修改为视讯里的参数后运行。不过所给实例里的两个参数并未给出，个人只能指定一个猜测的数值。

代码如下（来源：https://blog.csdn.net/weixin_43584807/article/details/105784040）

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

#Disable TF2 behavior
#tf.disable_v2_behavior()
#Disable executing eagerly
#tf.compat.v1.disable_eager_execution();
#check whether the executing eagerly is Enabled
tf.executing_eagerly();

#VALUE
training_epochs = 10
learning_rate = 0.02

#primary Code(s) starts here
num_points = 1000
def data():
    x_data = np.linspace(0,1,1000)
    np.random.seed(5)
    ## y=3.1234*x+2.98+噪声（噪声的维度要和x_data一致）
    y_data = 0.1 * x_data + 0.3 + np.random.randn(*x_data.shape) * 0.02
    return x_data,y_data

x_data,y_data = data()

w = tf.Variable(np.random.randn(),tf.float32)
##构建模型中的变量b，对应线性函数的截距
b = tf.Variable(0.0,tf.float32)
def y_function(x,w,b):
    return w * x + b
def loss(x,y,w,b):
    return tf.reduce_mean(tf.square(y_function(x,w,b)-y))
def grad(x,y,w,b):
    with tf.GradientTape() as tape:
        loss_ = loss(x,y,w,b)
    return tape.gradient(loss_,[w,b])

step = 0
loss_list = [] #List of loss value
display_step = 10
for epoch in range(training_epochs):
    for xs,ys in zip(x_data,y_data):
        loss_ = loss(xs,ys,w,b)
        loss_list.append(loss_)
        delta_w,delta_b = grad(xs,ys,w,b)
        change_w = delta_w * learning_rate
        change_b = delta_b * learning_rate
        w.assign_sub(change_w)
        b.assign_sub(change_b)
        step=step+1;
        if step%(2*display_step)==0:
            print("TE:",'%02d'%(epoch+1),"Step:%03d"%(step),"LOSS=%.6f"%(loss_))
    plt.plot(x_data,w.numpy()*x_data+b.numpy())
print('w:',w.numpy())
print('b:',b.numpy())
print('最小的损失：',min(loss_list).numpy())
plt.figure(figsize=(10,6))
plt.scatter(x_data,y_data,label='Orignal data')
plt.plot(x_data, 0.1 * x_data + 0.3, label ='Object Line',c='g')
plt.plot(x_data, x_data * w.numpy() + b.numpy(), label='Fitted Line',color='r')
plt.legend(loc=2)
plt.figure(figsize=(10,6))
plt.plot(loss_list)
plt.show()

运行结果：