tensorflow MNIST Convolutional Neural Network

tensorflow MNIST Convolutional Neural Network

MNIST CNN 包含的几个部分：

• Weight Initialization
• Convolution and Pooling
• Convolution layer
• Fully connected layer
• Readout Layer

直接上tensorflow 给的示例：

先读入数据：

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

import tensorflow as tf
import time
#Weight	Initialization
#现在还没有值，只是计算图中的节点，直到‘tf.global_variables_initializer()’才初始化
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)

#Convolution 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')

tf.nn.conv2d(input, filter, strides, padding, use_cudnn_on_gpu=None, name=None)

• 第一个参数input：指需要做卷积的输入图像，它要求是一个Tensor，具有[batch, in_height, in_width, in_channels]这样的shape，具体含义是[训练时一个batch的图片数量, 图片高度, 图片宽度, 图像通道数]，注意这是一个4维的Tensor，要求类型为float32和float64其中之一
• 第二个参数filter：相当于CNN中的卷积核，它要求是一个Tensor，具有[filter_height, filter_width, in_channels, out_channels]这样的shape，具体含义是[卷积核的高度，卷积核的宽度，图像通道数，卷积核个数]，要求类型与参数input相同，有一个地方需要注意，第三维in_channels，就是参数input的第四维
• 第三个参数strides：卷积时在图像每一维的步长，这是一个一维的向量，长度4
• 第四个参数padding：string类型的量，只能是"SAME","VALID"其中之一，这个值决定了不同的卷积方式（后面会介绍）
• 第五个参数：use_cudnn_on_gpu:bool类型，是否使用cudnn加速，默认为true

结果返回一个Tensor，这个输出，就是我们常说的feature map

#Input	(placeholder)
x = tf.placeholder(tf.float32,shape=[None,784])
y_ = tf.placeholder(tf.float32,shape=[None,10])

#Convolution layer
x_image = tf.reshape(x, [-1,28,28,1])

W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])

h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1) #维度为[-1,28,28,32]
h_pool1 = max_pool_2x2(h_conv1)#维度为[-1,14,14,32]

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) #维度为[-1,14,14,64]
h_pool2 = max_pool_2x2(h_conv2) #维度为[-1,7,7,64]

tf.reshape(x, [-1,28,28,1])其中的-1表示由后面的几个维度来确定，

例如：

t=[[1, 2], [3, 4], [5, 6], [7, 8]] ,那么t的维度是[4,2]

(1) reshape(t,[2,4])后，t为[[1, 2, 3, 4], [5, 6, 7, 8]]

(2) reshape(t,[-1,4])后，t同样为[[1, 2, 3, 4], [5, 6, 7, 8]]，所以这里的-1实际上为2。

#Fully 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)

#dropout
keep_prob = tf.placeholder(tf.float32)
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

训练时使用dropout，减少过拟合

#Readout	Layer
W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])
y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2

#Training and Evaluation
cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y_conv)) #Evaluation
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy) #optimizer
correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32)) #accuracy

sess=tf.InteractiveSession()
sess.run(tf.global_variables_initializer())
t1=time.time()
for i in range(4000):
    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}))
t2=time.time()
print(t2-t1)