"""
最终测试集结果0.9912
test accuracy 0.9912
"""
]import tensorflow as tf
from tensorflow.examples.tutorials.mnistimport input_data
def weight_variable(shape): # 产生正态分布的卷积核权重，若生成的值与均值的差值大于两倍的标准差，就重新生成
initial= tf.truncated_normal(shape,stddev=0.1)
return tf.Variable(initial)
def bias_variable(shape): # 产生偏差矩阵
initial= tf.constant(0.1,shape=shape)
return tf.Variable(initial)
def conv2d(x, W): # 卷积函数
return tf.nn.conv2d(x, W,strides=[1,1,1,1],padding='SAME') # 步长strides决定卷积的步伐，padding决定是否丢弃一部分，VALID丢弃，SAME不丢弃，不足之处补0
def max_pool_2x2(x): # 池化函数
return tf.nn.max_pool(x,ksize=[1,2,2,1],strides=[1,2,2,1],padding='SAME')
mnist= input_data.read_data_sets("MNIST_data/",one_hot=True) # 下载或读取数据集，one_hot编码
x= tf.placeholder(tf.float32,[None,784]) # n行784列的输入矩阵
W= tf.Variable(tf.zeros([784,10])) # 权重矩阵，784行10列，初始化为零
b= tf.Variable(tf.zeros([10])) # 偏置矩阵，10列的一个array
y_= tf.placeholder("float",[None,10]) # 训练集的标签
# 第一层卷积池化(采样)得到14*14*32
W_conv1= weight_variable([5,5,1,32])
b_conv1= bias_variable([32])
x_image= tf.reshape(x,[-1,28,28,1])
h_conv1= tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1) # 首先是卷积加上偏置，卷积后再使用激活函数进行映射（实际是卷积层每个feature maps上的神经元阵列进行卷积、映射）
h_pool1= max_pool_2x2(h_conv1) # 实际上是进行采样，减小数据量，提取主要特征
# 第二层卷积池化(采样)得到7*7*64
W_conv2= weight_variable([5,5,32,64])
b_conv2= bias_variable([64])
h_conv2= tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2= max_pool_2x2(h_conv2)
# 全连接层将7*7*64展开
W_fc1= weight_variable([7 * 7 * 64,1024]) # 7*7*64个输入，1024个神经元
b_fc1= bias_variable([1024])
h_pool2_flat= tf.reshape(h_pool2,[-1,7 * 7 * 64])
h_fc1= tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)
# Dropout，减少过拟合
keep_prob= tf.placeholder('float')
h_fc1_drop= tf.nn.dropout(h_fc1, keep_prob) # 一般用在全连接层,其含义是指，以keep_prob概率变为原来的1/keep_prob,以keep_prob概率变为0
# 输出层,输出到10个神经元，构造softmax regression
W_fc2= weight_variable([1024,10])
b_fc2= bias_variable([10])
y_conv= tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
# 训练和评估模型
cross_entropy= -tf.reduce_sum(y_*tf.log(y_conv)) # 计算交叉熵,用来衡量模型好坏
train_step= tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
correct_prediction= tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1)) # 正确的标签与预测的标签进行比对，确定正确率
accuracy= tf.reduce_mean(tf.cast(correct_prediction,"float")) # argmax当axis=0时返回每一列的最大值的位置索引,当axis=1时返回每一行中的最大值的位置索引
sess= tf.InteractiveSession() # 交互式环境比Session更加灵活
sess.run(tf.global_variables_initializer())
for iin range(20000):
batch= mnist.train.next_batch(50)
if i% 100 == 0:
train_accuracy= accuracy.eval(feed_dict={x: batch[0], y_: batch[1], keep_prob: 1})
print("step %d ,training accuracy %g " % (i, train_accuracy))
train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})
print("test accuracy %g" % accuracy.eval(feed_dict={x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))
