上一講筆者和大家一起學(xué)習(xí)了如何使用 Tensorflow 構(gòu)建一個(gè)卷積神經(jīng)網(wǎng)絡(luò)模型。本節(jié)我們將繼續(xù)利用 Tensorflow 的便捷性完成 mnist 手寫(xiě)數(shù)字?jǐn)?shù)據(jù)集的識(shí)別實(shí)戰(zhàn)。mnist 數(shù)據(jù)集是 Yann Lecun 大佬基于美國(guó)國(guó)家標(biāo)準(zhǔn)技術(shù)研究所構(gòu)建的一個(gè)研究深度學(xué)習(xí)的手寫(xiě)數(shù)字的數(shù)據(jù)集。mnist 由 70000 張不同人手寫(xiě)的 0-9 10個(gè)數(shù)字的灰度圖組成。本節(jié)筆者就和大家一起研究如何利用 Tensorflow 搭建一個(gè) CNN 模型來(lái)識(shí)別這些手寫(xiě)的數(shù)字。

數(shù)據(jù)導(dǎo)入

mnist 作為標(biāo)準(zhǔn)深度學(xué)習(xí)數(shù)據(jù)集，在各大深度學(xué)習(xí)開(kāi)源框架中都默認(rèn)有進(jìn)行封裝。所以我們直接從 Tensorflow 中導(dǎo)入相關(guān)的模塊即可：

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

快速搭建起一個(gè)簡(jiǎn)易神經(jīng)網(wǎng)絡(luò)模型

數(shù)據(jù)導(dǎo)入之后即可按照 Tensorflow 的范式創(chuàng)建相應(yīng)的 Tensor 變量然后創(chuàng)建會(huì)話：

# create the session
sess = tf.InteractiveSession()
# create variables and run the session
x = tf.placeholder('float', shape=[None, 784])
y_ = tf.placeholder('float', shape=[None, 10])
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))
sess.run(tf.global_variables_initializer())

定義前向傳播過(guò)程和損失函數(shù)：

#definethenetandlossfunctiony=tf.nn.softmax(tf.matmul(x,W)+b)
cross_entropy=-tf.reduce_sum(y_*tf.log(y))

進(jìn)行模型訓(xùn)練：

# train the model
train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)
for i in range(1000):
 batch = mnist.train.next_batch(50)
 train_step.run(feed_dict={x: batch[0], y_: batch[1]})

使用訓(xùn)練好的模型對(duì)測(cè)試集進(jìn)行預(yù)測(cè)：

# evaluate the model
correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))
print(accuracy.eval(feed_dict={x: mnist.test.images, y_: mnist.test.labels}))

預(yù)測(cè)準(zhǔn)確率為 0.9，雖然說(shuō)也是一個(gè)很高的準(zhǔn)確率了，但對(duì)于 mnist 這種標(biāo)準(zhǔn)數(shù)據(jù)集來(lái)說(shuō)，這樣的結(jié)果還有很大的提升空間。所以我們繼續(xù)優(yōu)化模型結(jié)構(gòu)，為模型添加卷積結(jié)構(gòu)。

搭建卷積神經(jīng)網(wǎng)絡(luò)模型

定義初始化模型權(quán)重函數(shù)：

# initilize the weight
def weight_variable(shape):
  initial = tf.truncated_normal(shape, stddev=0.1) 
  return tf.Variable(initial)
  
def bias_variable(shape):
  initial = tf.constant(0.1, shape=shape) 
  return tf.Variable(initial)

定義卷積和池化函數(shù)：

# convolutional and pooling
def conv2d(x, W): 
  return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

def max_pool_2x2(x): 
  return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],
            strides=[1, 2, 2, 1], padding='SAME')

搭建第一層卷積：

# the first convolution layer
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)
h_pool1 = max_pool_2x2(h_conv1)

搭建第二層卷積：

# the second convolution layer
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)

搭建全連接層：

# dense layer/full_connected layer
W_fc1 = weight_variable([7 * 7 * 64, 1024])
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)

設(shè)置 dropout 防止過(guò)擬合：

# dropout to prevent overfitting
keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

對(duì)輸出層定義 softmax ：

# model output
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)

訓(xùn)練模型并進(jìn)行預(yù)測(cè)：

# model trainning and evaluating
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))
  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}))

部分迭代過(guò)程和預(yù)測(cè)結(jié)果如下：

經(jīng)過(guò)添加兩層卷積之后我們的模型預(yù)測(cè)準(zhǔn)確率達(dá)到了 0.9931，模型訓(xùn)練的算是比較好了