莫烦 tensorflow 笔记 （一）搭建神经网络

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Wed Aug 16 08:56:03 2017

@author: xiaolian
"""
# add layers

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

# inputs 输入值
# in_size 输入的大小 
# out_size ‘输出的大小
# activation_function 激励函数

def add_layer(inputs, in_size, out_size, activation_function = None):
    Weights = tf.Variable(tf.random_normal([in_size, out_size]))

    #tf.cast(Weights, tf.float64)  
    #tf.cast(inputs, tf.float64)

    biases = tf.Variable(tf.zeros([1, out_size]) + 0.1)
    #tf.cast(biases, tf.float64)

    z =  tf.matmul(inputs, Weights) + biases
    if activation_function is None:
        outputs = z
    else:
        outputs = activation_function(z)
    return outputs



x_data = np.linspace(-1, 1, 300)[:,np.newaxis].astype(np.float32)
noise = np.random.normal(0, 0.05, x_data.shape).astype(np.float32)
y_data = (np.square(x_data) - 0.5 + noise).astype(np.float32)

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)

# add output layer
prediction = add_layer(l1, 10, 1, activation_function = 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 variable
init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init)

# visualization
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
ax.scatter(x_data, y_data)
plt.ion()

for i in range(1000):
    sess.run(train_step, feed_dict = {xs:x_data, ys:y_data})
    if i == 999:
        print(sess.run(loss, feed_dict = {xs:x_data, ys:y_data}))
        prediction_value = sess.run(prediction, feed_dict = {xs:x_data})
        line = ax.plot(x_data, prediction_value, 'g', lw = 5)
plt.show()

输出：