自建神经网络与结果可视化

来自周莫烦Tensorflow教学视频，主要内容有添加神经层，和训练结果可视化。

youtube地址： https://www.youtube.com/watch?v=nhn8B0pM9ls&list=PLXO45tsB95cKI5AIlf5TxxFPzb-0zeVZ8&index=16

import tensorflow as tf
import numpy as np
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
import matplotlib.pyplot as plt
def add_layer(inputs, in_size, out_size, activation_function = None):
    #矩阵
    Weights = tf.Variable(tf.random_normal([in_size,out_size]))
    bias = tf.Variable(tf.zeros([1,out_size]) + 0.1)
    Wx_plus_b = tf.matmul(tf.cast(inputs,tf.float32),Weights) + bias
    if activation_function is None:
        outputs = Wx_plus_b
    else:
        outputs = activation_function(Wx_plus_b)
    return outputs

if 'session' in locals() and session is not None:
    print('Close interactive session')
    session.close()

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])

l1 = add_layer(x_data,1,10,activation_function=tf.nn.relu)
prediction = add_layer(l1,10,1,activation_function=None)

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.global_variables_initializer()
sess = tf.Session()
sess.run(init)

fig = plt.figure()
ax = fig.add_subplot(1,1,1) #连续性，编号
ax.scatter(x_data,y_data)
plt.ion() #不暂停主程序
plt.show()

for i in range(1000):
    sess.run(train_step,feed_dict={xs:x_data,ys:y_data})
    if i % 50 == 0:
        try:
            ax.lines.remove(lines[0])
        except Exception:
            pass
        #print(sess.run(loss,feed_dict={xs:x_data,ys:y_data}))
        prediction_value = sess.run(prediction, feed_dict={xs:x_data})
        lines = ax.plot(x_data,prediction_value,'r-',lw=5)
        plt.pause(0.1)

 结果显示：

 

特别注意的是其中有三行代码为：

if 'session' in locals() and session is not None:
    print('Close interactive session')
    session.close()

是因为出现了“InternalError (see above for traceback): Blas GEMM launch failed : a.shape=(300, 1), b.shape=(1, 10), m=300, n=10, k=1”的错误，stackoverflow上给的解答大致意思为session已在其他进程中出现，所以先关闭session.