import matplotlib.pyplot as plt
  plt.ion()  # 打開(kāi)交互模式
  # 同時(shí)打開(kāi)兩個(gè)窗口顯示圖片
  plt.figure() #圖片一
  plt.imshow(i1)
  plt.figure()  #圖片二
  plt.imshow(i2)
  # 顯示前關(guān)掉交互模式
  plt.ioff()
  plt.show()

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
 
def add_layer(inputs,in_size,out_size,activation_funiction=None):
 
  Weights = tf.Variable(tf.random_normal([in_size,out_size]))
  biases = tf.Variable(tf.zeros([1,out_size]) +0.1)
  Wx_plus_b = tf.matmul(inputs,Weights)+biases
  if activation_funiction is None:
    outputs = Wx_plus_b
  else:
    outputs = activation_funiction(Wx_plus_b)
  return outputs
 
x_data = np.linspace(-1,1,300)[:,np.newaxis]
noise = np.random.normal(0,0.05,x_data.shape)
y_data = np.square(x_data)-0.5 +noise
 
xs = tf.placeholder(tf.float32,[None,1])  
ys = tf.placeholder(tf.float32,[None,1])
 
#add hidden layer
l1 = add_layer(xs,1,10,activation_funiction=tf.nn.relu)
#add output layer
prediction = add_layer(l1,10,1,activation_funiction=None)
 
#the error between prediction and real data
loss = tf.reduce_mean(tf.reduce_sum(tf.square(ys - prediction),reduction_indices=[1]))
train_step = tf.train.GradientDescentOptimizer(0.1).minimize(loss)
 
init =tf.initialize_all_variables()
 
with tf.Session() as sess:
  sess.run(init)
 
  fig = plt.figure()
  ax = fig.add_subplot(1,1,1)
  ax.scatter(x_data,y_data)
  plt.ion()  #將畫(huà)圖模式改為交互模式
 
  for i in range(1000):
    sess.run(train_step,feed_dict={xs:x_data,ys:y_data})
    if i%50 ==0:
      plt.pause(0.1)
      try:
        ax.lines.remove(lines[0])
      except Exception:
        pass
      prediction_value = sess.run(prediction,feed_dict={xs:x_data})
      lines = ax.plot(x_data,prediction_value,'r-',lw=5)
 
 
      print(sess.run(loss,feed_dict={xs:x_data,ys:y_data}))
 
  plt.ioff()
  plt.show()