tensorflow-线性回归

import tensorflow as tf
import numpy as np

"""
x_data=np.random.rand(200,4)
w_data=np.array([1,-2,3,-4])
w_data=w_data.reshape(4,1)
b_data=np.array([3.0])
y_data=np.matmul(x_data,w_data)+b_data
"""
#---------------------------------------------------------生成数据
weights=tf.constant([1.,-2.,3.,-4.])
weights=tf.reshape(weights,[4,1])
bias=tf.constant(3.0)

inputData=tf.Variable(tf.random_normal([200,4]))
outputData=tf.matmul(inputData,weights)+bias


#---------------------------------------------------------定义图
x=tf.placeholder(tf.float32,shape=[200,4])
y=tf.placeholder(tf.float32,shape=[200,1])

w=tf.Variable(tf.random_normal([4,1]),name="weight")
b=tf.Variable(1.0,name="bias")

y1=tf.matmul(x,w)+b
loss=tf.reduce_mean(tf.square(y1-y))
optimizer=tf.train.GradientDescentOptimizer(learning_rate=0.01)
train_step=optimizer.minimize(loss)

init=tf.global_variables_initializer()

with tf.Session() as sess:
    sess.run(init)
    #print(sess.run(inputData))
    #print(sess.run(outputData))
    x_data =inputData.eval(session=sess)
    y_data= outputData.eval(session=sess)

    for i in range(10000):
        sess.run(train_step, feed_dict={x: x_data, y: y_data})
        cost=sess.run(loss,feed_dict={x:x_data,y:y_data})
        if i % 100 == 0:
            #打印拟合出来的值
            print("Epoch:", '%04d' % (i + 1), "loss=", cost, \
                  "W=", sess.run(w), "b=", sess.run(b))