TensorFlow从入门到理解（二）：你的第一个神经网络

运行代码：

from __future__ import print_function
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

# 神经层函数
def add_layer(inputs, in_size, out_size, activation_function=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_function is None:
        outputs = Wx_plus_b
    else:
        outputs = activation_function(Wx_plus_b)
    return outputs

# 导入数据
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

# 利用占位符定义我们所需的神经网络输入
xs = tf.placeholder(tf.float32, [None, 1])
ys = tf.placeholder(tf.float32, [None, 1])

# 定义隐藏层
l1 = add_layer(xs, 1, 10, activation_function=tf.nn.relu)

# 定义输出层
prediction = add_layer(l1, 10, 1, activation_function=None)

# 计算误差和提供准确率
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()
# 输出结果
sess = tf.Session()
sess.run(init)

# matplotlib可视化
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.scatter(x_data, y_data)
plt.ion()
plt.show()

# 机器学习，学习1000次
for i in range(1000):
    # 每50步输出学习误差
    sess.run(train_step, feed_dict={xs: x_data, ys: y_data})
    if i % 50 == 0:
        # 可视化结果和改进
        try:
            ax.lines.remove(lines[0])
        except Exception:
            pass
        prediction_value = sess.run(prediction, feed_dict={xs: x_data})
        # 用红色和宽度为5的线来显示预测结果，并暂停0.1秒
        lines = ax.plot(x_data, prediction_value, 'r-', lw=5)
        plt.pause(1)

运行结果：