tensorflow mnist数据集 普通神经网络

　　今天写一个mnist数据集的神经网络识别 

　　mnist数据集是机器学习的Hello World ，这篇博客用最简单的神经网络classfy 数据集中的数字

1.import tensorflow 然后载入数据集，在tensorflow中google已经帮我们封装好了这个数据集，直接  从tensorflow.example.tutorial.mnist中 import就行。然后input_data中读入数据，在hot_hot参数中设置True。他会帮我们自动编码成‘one-hot’

import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
# number 1 to 10 data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)

2.这里我们写一个add_layer的函数，增加一个隐藏层，参数有inputs：输入数据， in_size：输入数据的大小， out_size：输出数据的大小, activate_function:激活函数，初始值设为None，当传入激活函数时，add_layer会输出经过激活函数之后的值

def add_layer(inputs, in_size, out_size, activate_function=None):
    Weight = 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, Weight) + biases
    if activate_function is None:
        outputs = Wx_plus_b
    else:
        outputs = activate_function(Wx_plus_b)
    return outputs

 3.新建两个占位符 ys（labels) 它的大小是10，也就是数字0到9的one-hot编码  xs(features) 是每幅图像的像素点（28*28）

ys = tf.placeholder(tf.int64, [None, 10])
xs = tf.placeholder(tf.float32, [None, 784])

4.logits： logits就是神经网络的输出值。没有经过softmax函数缩放的值

logits= add_layer(xs, 784, 10, activate_function=None)

5.计算损失函数。 这里我用tf.nn.softmax_cross_entropy_with_logtis,这是一个tensorflow封装好了的函数，这边的可以直接把logtis和labels放进去，就可以直接得到经过soft_max后的logits和labels的交叉熵loss

关于几个不同的损失函数，可以看我另一篇博客整理的内容https://www.cnblogs.com/francischeng/p/9836341.html

loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=ys))

6.生成优化器，这里我选择了GrandientDescent方法进行优化 ，学习率(learning rate)设置成了0.5

train_step = tf.train.GradientDescentOptimizer(0.5).minimize(loss)

7.def计算准确度函数 :

　这边传入的参数v_xs和v_ys将会是整个数据集的数据，会计算模型对于整个数据集的准确度

def compute_accuracy(v_xs, v_ys):
    y_pre = sess.run(logits, feed_dict={xs: v_xs})
    correct_prediction = tf.equal(tf.arg_max(y_pre, 1), tf.arg_max(v_ys, 1))
    accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
    result = sess.run(accuracy, feed_dict={xs:v_xs, ys:v_ys})
    return result

8. 跑起来之前的准备 

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

9.跑起来

for i in range(1000):
    batch_xs, batch_ys = mnist.train.next_batch(100)
    sess.run(train_step, feed_dict={xs:batch_xs, ys:batch_ys})
    if i%50 == 0:
        print(compute_accuracy(mnist.test.images, mnist.test.labels))

完整代码1

import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)


def add_layer(inputs, in_size, out_size, activate_function=None):
    Weight = 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, Weight) + biases
    if activate_function is None:
        outputs = Wx_plus_b
    else:
        outputs = activate_function(Wx_plus_b)
    return outputs
    
def compute_accuracy(v_xs, v_ys):
    y_pre = sess.run(logits, feed_dict={xs: v_xs})
    correct_prediction = tf.equal(tf.argmax(y_pre, 1), tf.argmax(v_ys, 1))
    accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
    result = sess.run(accuracy, feed_dict={xs:v_xs, ys:v_ys})
    return result

ys = tf.placeholder(tf.int64, [None, 10])
xs = tf.placeholder(tf.float32, [None, 784])


logits= add_layer(xs, 784, 10, activate_function=None)

loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=ys))

train_step = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
init = tf.global_variables_initializer()



sess = tf.Session()
init = tf.global_variables_initializer()
sess.run(init)
for i in range(1000):
    batch_xs, batch_ys = mnist.train.next_batch(100)
    sess.run(train_step, feed_dict={xs:batch_xs, ys:batch_ys})
    if i%50 == 0:
        print(compute_accuracy(mnist.test.images, mnist.test.labels))

完整代码2：

这个是没有用tf.nn.softmax_cross_entropy_with_logits， 手动添加了soft_max并计算了交叉熵

import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data

mnist = input_data.read_data_sets('MNIST_data', one_hot=True)


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


def compute_accuracy(v_xs, v_ys):
    global prediction
    y_pre = sess.run(prediction, feed_dict={xs: v_xs})
    correct_prediction = tf.equal(tf.argmax(y_pre, 1), tf.argmax(v_ys, 1))
    accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
    result = sess.run(accuracy, feed_dict={xs: v_xs, ys: v_ys})
    return result


xs = tf.placeholder(tf.float32, [None, 784])  # 28x28
ys = tf.placeholder(tf.float32, [None, 10])

prediction = add_layer(xs, 784, 10, activation_function=tf.nn.softmax)


cross_entropy = tf.reduce_mean(-tf.reduce_sum(ys * tf.log(prediction),
                                              reduction_indices=[1]))
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)

sess = tf.Session()


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

for i in range(1000):
    batch_xs, batch_ys = mnist.train.next_batch(100)
    sess.run(train_step, feed_dict={xs: batch_xs, ys: batch_ys})
    if i % 50 == 0:
        print(compute_accuracy(
            mnist.test.images, mnist.test.labels))