由淺入深之Tensorflow(1)----linear_regression實現

　　Tensorflow是目前非常流行的deeplearning框架，學習Tensorflow最好的方法是github上的tf項目https://github.com/tensorflow/tensorflow

或者閱讀極客學院主導翻譯的中文教程http://wiki.jikexueyuan.com/project/tensorflow-zh/how_tos/reading_data.html 。

　　此處對tensorflow的基本語法不予贅述，直接貼上源碼：

import numpy as np
import tensorflow as tf

#准備數據
trainX = np.linspace(-1, 1, 101)
trainY = 2 * trainX + np.random.randn(*trainX.shape) * 0.33

#定義模型
def model(X, w):
    return tf.mul(X, w)

#初始化數據流圖
X = tf.placeholder('float')
Y = tf.placeholder('float')

w = tf.Variable(0.0, name = 'weights')

y_ = model(X, w)

#評估模型
cost = tf.square(Y - y_)
train_op = tf.train.GradientDescentOptimizer(0.01).minimize(cost)

sess = tf.InteractiveSession()
init = tf.initialize_all_variables()
#訓練
sess.run(init)
for i in range(100):
    for (x, y) in zip(trainX, trainY):
        sess.run(train_op, feed_dict = {X: x, Y: y})
print sess.run(w)

sess.close()