卷积多层实现手写数字识别

import tensorflow as tf
import random
import matplotlib.pyplot as plt
from tensorflow.examples.tutorials.mnist import input_data

tf.set_random_seed(777)

mnist = input_data.read_data_sets(‘MNIST_data’, one_hot=True)

alpha = 0.001
training_epoch = 15
batch_size = 100

keep_pre = tf.placeholder(tf.float32)

初始化x, y占位符

X = tf.placeholder(tf.float32, [None, 784])
X_img = tf.reshape(X, [-1, 28, 28, 1])
Y = tf.placeholder(tf.float32, [None, 10])

第一层卷积

W1 = tf.Variable(tf.random_normal([3, 3, 1, 32], stddev=0.01))
L1 = tf.nn.conv2d(X_img, W1, strides=[1, 1, 1, 1], padding=‘SAME’)
L1 = tf.nn.relu(L1)
L1 = tf.nn.max_pool(L1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding=‘SAME’)
L1 = tf.nn.dropout(L1, keep_prob=keep_pre)

第二层卷积

W2 = tf.Variable(tf.random_normal([3, 3, 32, 64], stddev=0.01))
L2 = tf.nn.conv2d(L1, W2, strides=[1, 1, 1, 1], padding=‘SAME’)
L2 = tf.nn.relu(L2)
L2 = tf.nn.max_pool(L2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding=‘SAME’)
L2 = tf.nn.dropout(L2, keep_prob=keep_pre)

第三层卷积

W3 = tf.Variable(tf.random_normal([3, 3, 64, 128], stddev=0.01))
L3 = tf.nn.conv2d(L2, W3, strides=[1, 1, 1, 1], padding=‘SAME’)
L3 = tf.nn.relu(L3)
L3 = tf.nn.max_pool(L3, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding=‘SAME’)
L3 = tf.nn.dropout(L3, keep_prob=keep_pre)
L3_flat = tf.reshape(L3, [-1, 12844])

全连接（第一层）

W4 = tf.get_variable(‘W4’, shape=[12844, 625])
b4 = tf.Variable(tf.random_normal([625]))
L4 = tf.nn.relu(tf.matmul(L3_flat, W4)+b4)
L4 = tf.nn.dropout(L4, keep_prob=keep_pre)

最后一层（分类层）

W5 = tf.get_variable(‘W5’, shape=[625, 10], initializer=tf.contrib.layers.xavier_initializer())
b5 = tf.Variable(tf.random_normal([10]))
logdits = tf.matmul(L4, W5)+b5

代价

cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logdits, labels=Y))

精确率

accuracy = tf.reduce_mean(tf.cast(tf.equal(tf.argmax(logdits, 1), tf.argmax(Y, 1)), tf.float32))

优化器

op = tf.train.AdamOptimizer(learning_rate=alpha).minimize(cost)

执行会话

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

for i in range(training_epoch):
avg_cost = 0
m = int(mnist.train.num_examples/batch_size)
for j in range(m):
xdata, ydata = mnist.train.next_batch(batch_size)
cost_var, _ = sess.run([cost, op], feed_dict={X: xdata, Y: ydata, keep_pre: 0.7})
avg_cost += cost_var/m
print(avg_cost)
print(sess.run(accuracy, feed_dict={X: mnist.test.images, Y:mnist.test.labels, keep_pre: 1}))

r = random.randint(0, mnist.test.num_examples-1)
print(‘标签’, sess.run(tf.argmax(mnist.test.labels[r:r+1], 1)))
print(‘预测值’, sess.run(tf.argmax(logdits, 1), feed_dict={X:mnist.test.images[r:r+1], keep_pre: 1}))

plt.imshow(mnist.test.images[r:r+1].reshape(28, 28), cmap=‘Greys’, interpolation=‘nearest’)
plt.show()