tensorflow实战2：实现mnist 单层cnn分类

1、数据处理

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

np.random.seed(20160703)
tf.set_random_seed(20160703)
mnist = input_data.read_data_sets("/tmp/data/", one_hot=True)


x = tf.placeholder(tf.float32, [None, 784])
x_image = tf.reshape(x, [-1,28,28,1])

2、卷积层的设计
第一层卷积层卷积核数量为16，由于通道为1，卷积核为5*5，则每个卷积核的参数w的维度为（5,5,1），因为有16个卷积核，则w为（5,5,1,16）的维度。strides参数官方解释
strides: A list of ints that has length >= 4. The stride of the sliding window for each dimension of the input tensor.
首先要求 strides 为长度不小于 4 的整数构成的 list，strides参数表示的是滑窗在输入张量各个维度上的移动步长。
池化层 ksize 2表示在图像的高跟宽做池化
strides则跟卷积层类似

num_filters = 16
W_conv = tf.Variable(tf.truncated_normal([5,5,1,num_filters],
                                         stddev=0.1))
h_conv = tf.nn.conv2d(x_image, W_conv,
                      strides=[1,1,1,1], padding='SAME')
h_pool =tf.nn.max_pool(h_conv, ksize=[1,2,2,1],
                       strides=[1,2,2,1], padding='SAME')

3、flatten

池化之后将池化层flatten
即打成14*14*16的一维向量
后面加一层relu再输出softmax搞定

h_pool_flat = tf.reshape(h_pool, [-1, 14*14*num_filters])

num_units1 = 14*14*num_filters
num_units2 = 1024

w2 = tf.Variable(tf.truncated_normal([num_units1, num_units2]))
b2 = tf.Variable(tf.zeros([num_units2]))
hidden2 = tf.nn.relu(tf.matmul(h_pool_flat, w2) + b2)

w0 = tf.Variable(tf.zeros([num_units2, 10]))
b0 = tf.Variable(tf.zeros([10]))
p = tf.nn.softmax(tf.matmul(hidden2, w0) + b0)

4、启动session

t = tf.placeholder(tf.float32, [None, 10])
loss = -tf.reduce_sum(t * tf.log(p))
train_step = tf.train.AdamOptimizer(0.0005).minimize(loss)
correct_prediction = tf.equal(tf.argmax(p, 1), tf.argmax(t, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))


sess = tf.InteractiveSession()
sess.run(tf.initialize_all_variables())
saver = tf.train.Saver()


i = 0
for _ in range(4000):
    i += 1
    batch_xs, batch_ts = mnist.train.next_batch(100)
    sess.run(train_step, feed_dict={x: batch_xs, t: batch_ts})
    if i % 100 == 0:
        loss_val, acc_val = sess.run([loss, accuracy],
            feed_dict={x:mnist.test.images, t:mnist.test.labels})
        print ('Step: %d, Loss: %f, Accuracy: %f'
               % (i, loss_val, acc_val))
        saver.save(sess, 'mdc_session', global_step=i)