tensorflow (4) mnist·CNN

0.文件目录如图
1.CNN实现如下，每行代码都有注释
import tensorflow as tf
import input_data
#卷积网络
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)   #载入数据
sess = tf.InteractiveSession()  #创建会话

x = tf.placeholder("float", shape=[None, 784]) #创建变量x,训练时输入数据，None一次输入的样本数量，784是照片转换的数组
y_ = tf.placeholder("float", shape=[None, 10]) #创建变量y_，训练时输出数据  ，有十个数字，对应输出数组长度为10

def weight_variable(shape): #权重，shape待输入
  initial = tf.truncated_normal(shape, stddev=0.1)   #tf.truncated_normal(shape, stddev=0.1),shape表示生成张量的维度，mean是均值，stddev是标准差，truncated为截断正态分布
                                                     #生成[n,n]张量，一个方差为零的正态分布，这样防止梯度为零，应为激励函数为RELU函数，初始权重为零的话，节点输出均为零
  return tf.Variable(initial)    #返回权重变量初始值

def bias_variable(shape):  #偏置
  initial = tf.constant(0.1, shape=shape)  #偏置为常量
  return tf.Variable(initial)


def conv2d(x, W):   ####卷积
  return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')  #[1，stride，stride，1]图像只有两维，中间两个参数是二维方向上的步长 ‘VALID’ 'SAME'（卷积核可以停留在图像边缘）

def max_pool_2x2(x):  ####池化
  return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],strides=[1, 2, 2, 1], padding='SAME')  #ksize：池化窗口的大小，取一个四维向量，一般是[1, height, width, 1]，因为我们不想在batch和channels上做池化，所以这两个维度设为了1

###第一层卷积
W_conv1 = weight_variable([5, 5, 1, 32])  #参数为shape，5*5，单通道输入，32通道特征输出
b_conv1 = bias_variable([32])  #参数为shape，32通道特征输出，故b的shape为长度32  [ 0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1 0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1  0.1 0.1  0.1]


x_image = tf.reshape(x, [-1,28,28,1])  #[-1,28,28,1] batch数量不限，28*28像素的图片  深度为1
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)  #激活函数
h_pool1 = max_pool_2x2(h_conv1)                    #池化

###第二层卷积
W_conv2 = weight_variable([5, 5, 32, 64])   #5*5核，32输入，64输出
b_conv2 = bias_variable([64])

h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)

###密集连接层
W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])

h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])   #把最后一层池化层转为  7*7*64列的矩阵
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

###Dropout
keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

###输出层
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])

###训练和评估模型
y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
cross_entropy = -tf.reduce_sum(y_*tf.log(y_conv))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))
sess.run(tf.initialize_all_variables())
for i in 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: 1.0})
    print("step %d, training accuracy %g"%(i, train_accuracy))
    # print(sess.run(y))
  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}))
tensorflow (4) mnist·CNN

猜你喜欢
