结果可视化
1 import tensorflow as tf 2 import numpy as np 3 import matplotlib.pyplot as plt 4 def add_layer(inputs, in_size, out_size,activation_function=None): 5 Weights = tf.Variable(tf.random_normal([in_size, out_size])) 6 biases = tf.Variable(tf.zeros([1, out_size]) + 0.1) 7 Wx_plus_b = tf.matmul(inputs, Weights) + biases 8 if activation_function is None: 9 outputs = Wx_plus_b 10 else: 11 outputs = activation_function(Wx_plus_b) 12 return outputs 13 14 x_data=np.linspace(-1,1,300,dtype=np.float32)[:,np.newaxis] 15 noise = np.random.normal(0, 0.05, x_data.shape).astype(np.float32) 16 y_data=np.square(x_data)-0.5+noise 17 xs=tf.placeholder(tf.float32,[None,1],name='x_input') 18 ys=tf.placeholder(tf.float32,[None,1],name='y_input') 19 20 l1=add_layer(xs,1,10,activation_function=tf.nn.relu) #隐藏层 21 prediction=add_layer(l1,10,1,activation_function=None) #输出层 22 loss=tf.reduce_mean(tf.reduce_sum(tf.square(ys-prediction), 23 reduction_indices=[1])) 24 train_step = tf.train.GradientDescentOptimizer(0.1).minimize(loss) 25 init = tf.global_variables_initializer() 26 sess = tf.Session() 27 sess.run(init) 28 fig = plt.figure() #生成图片框架 29 ax=fig.add_subplot(1,1,1) #连续性的画图 30 ax.scatter(x_data,y_data) #用点的形式把真实的数据画出来 31 plt.ion() #用于连续显示，不会show一下就停止显示 32 plt.show() 33 for i in range(1000): 34 sess.run(train_step,feed_dict={xs:x_data,ys:y_data}) 35 if i%50==0: 36 print(sess.run(loss,feed_dict={xs:x_data,ys:y_data})) 37 try: 38 ax.lines.remove(lines[0]) #在图片中去除第一条线 39 except Exception: 40 pass 41 prediction_value = sess.run(prediction,feed_dict={xs:x_data}) 42 lines=ax.plot(x_data,prediction_value,'r-',lw=5) #红色，宽度为5的线,x,y轴的数据plot上去
43 plt.pause(0.1) #暂停0.1s
