完整程式碼

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

#樣本資料
x_train = np.linspace(-1,1,300)[:,np.newaxis]
noise = np.random.normal(0, 0.1, x_train.shape)
y_train = x_train * 3 + noise + 0.8

#線性模型
W = tf.Variable([0.1],dtype = tf.float32)
b = tf.Variable([0.1],dtype = tf.float32)
x = tf.placeholder(tf.float32)
line_model = W * x + b

#損失模型
y = tf.placeholder(tf.float32)
loss = tf.reduce_sum(tf.square(line_model - y))

#建立優化器
optimizer = tf.train.GradientDescentOptimizer(0.001)
train = optimizer.minimize(loss)

#初始化變數
init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init)

# 繪製樣本資料
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.scatter(x_train, y_train)
plt.ion()
plt.show()
plt.pause(3)


#訓練100次
for i in range(100):
    #每隔10次列印1次成果
    if i % 10 == 0:
        print(i)
        print('W:%s  b:%s' % (sess.run(W),sess.run(b)))
        print('loss:%s' % (sess.run(loss,{x:x_train,y:y_train})))
    sess.run(train,{x:x_train,y:y_train})

print('---')
print('W:%s  b:%s' % (sess.run(W),sess.run(b)))
print('loss:%s' % (sess.run(loss,{x:x_train,y:y_train})))
 

樣本訓練資料分佈如下

輸出結果如下

結論

通過列印結果可以看到W已經非常接近初始給定的3，b也非常接近給定的0.8 (誤差不可避免)