Tensorflow - Tutorial (1) : 线性回归

使用tensorflow做线性回归
程序运行结果如下图：

实验代码:

#!/usr/bin/env pythonimport tensorflow as tfimport numpy as npimport matplotlib.pyplot as plt# 生成101组二维训练数据 trX = np.linspace(-1, 1, 101) #创建-1到1之间的一维等差数组trY = 2 * trX + np.random.randn(*trX.shape) * 0.33 # 目标函数为y=2x，并施加一个高斯分布的噪音X = tf.placeholder("float") #创建占位符Y = tf.placeholder("float")def model(X, w):    return tf.mul(X, w) # 线性回归模型：x与w的内积w = tf.Variable(0.0, name="weights") # 为模型权重W创建共享变量y_model = model(X, w) cost = tf.square(Y - y_model) # 定义平方损失函数train_op = tf.train.GradientDescentOptimizer(0.01).minimize(cost) #使用步长为0.01的梯度下降算法，最小化损失函数cost# 创建会话with tf.Session() as sess:    # 初始化变量W    tf.initialize_all_variables().run()    for i in range(100):        for (x, y) in zip(trX, trY):            sess.run(train_op, feed_dict={X: x, Y: y})    weight = sess.run(w)    print(weight)   #输出训练完成的权重w，结果应接近2#画图fig = plt.figure()ax1 = fig.add_subplot(111)xx = np.linspace(-1, 1, 101) yy = xx*weightax1.scatter(trX ,trY)plt.plot(xx,yy,"r")plt.show()