深度学习之TensorFlow进阶教程一

最新推荐文章于 2024-01-18 02:09:08 发布

原创最新推荐文章于 2024-01-18 02:09:08 发布 · 539 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#tensorflow #深度学习 #python

tensorflow 专栏收录该内容

11 篇文章

订阅专栏

本文介绍如何利用TensorFlow和Python实现简单的数据拟合任务。通过加载外部数据集并构建神经网络模型来预测输出值，同时展示了如何训练模型及可视化训练过程。

声明：本人所用操作系统为Ubuntu kylin16.04LST，编辑器为Vim，Tensorflow版本为0.8，python版本为2.7。

一.读取数据

数据
数据文件：

np_x1.txt
np_y1.txt
链接: https://pan.baidu.com/s/1c1M1PQC 密码: ysg9
读数据

#!/usr/bin/python
# coding=utf-8

import numpy as np

# Make up some real data
f_x = open('np_x1.txt')
x_data= np.loadtxt(f_x,unpack='true')
x_data.shape = 400,1
f_x.close()

f_y = open('np_y1.txt')
y_data = np.loadtxt(f_y,unpack='true')
y_data.shape = 400,1
f_y.close()

二.模型的建立与训练

#!/usr/bin/python
# coding=utf-8

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

def net_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

# read data
f_x = open('np_x1.txt')
x_data= np.loadtxt(f_x,unpack='true')
x_data.shape = 400,1
f_x.close()

f_y = open('np_y1.txt')
y_data = np.loadtxt(f_y,unpack='true')
y_data.shape = 400,1
f_y.close()


# define placeholder for inputs to network
xs = tf.placeholder(tf.float32, [None, 1])
ys = tf.placeholder(tf.float32, [None, 1])
# add hidden layer
l1 = net_layer(xs, 1, 10, activation_function=tf.nn.relu)
# add output layer
prediction = net_layer(l1, 10, 1, activation_function=None)

# the error between prediciton 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()
sess= tf.Session()
sess.run(init)

# plot data
fig = plt.figure()
ax = fig.add_subplot(1,1,1) #图像为１行１列，第一个
p = ax.scatter(x_data, y_data) #散点图
plt.ion() #打开绘图交互模式
plt.show()


for i in range(10000):
    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
        prediction_value = sess.run(prediction, feed_dict={xs: x_data})
        # plot the prediction
        lines = ax.plot(x_data, prediction_value, 'r-', lw=5)
        plt.pause(1)