Tensorflow实现进阶的卷积网络

此部分代码来自于《Tensorflow实战》5.3节

#前两行代码可以跳过，自行去
https://github.com/tensorflow/models/tree/master/tutorials/image/cifar10
中将cifar10.py和cifar10_input.py下载下来
'''
git clone https://github.com/tensorflow/models.git
cd models/tutorials/image/cifar10
'''

import cifar10,cifar10_input
import tensorflow as tf
import numpy as np
import time

max_steps = 3000
batch_size = 128
data_dir = 'D:/Sublime Text 3/cifar10_data/cifar-10-batches-bin'#此处存放的路径我写的绝对路径  书中写的/tmp应该是IDLE的安装位置

def variable_with_weight_loss(shape, stddev, wl):
	var = tf.Variable(tf.truncated_normal(shape, stddev = stddev))
	if wl is not None:
		weight_loss = tf.multiply(tf.nn.l2_loss(var), wl, name = 'weight_loss')
		tf.add_to_collection('losses', weight_loss)
	return var
#把cifar10的数据解压到data_dir中，然后将下一行代码注释掉，取消运行
#cifar10.maybe_download_and_extract()

images_train, labels_train = cifar10_input.distorted_inputs(data_dir = data_dir, batch_size = batch_size)
#这行代码在Tensorflow 0.12版中会报错【TypeError: strided_slice() missing 1 required positional argument: 'strides'】我将版本升至1.0后就没有报错了

images_test, labels_test = cifar10_input.inputs(eval_data = True, data_dir = data_dir, batch_size = batch_size)

image_holder = tf.placeholder(tf.float32, [batch_size, 24, 24, 3])
label_holder = tf.placeholder(tf.int32, [batch_size])

weight1 = variable_with_weight_loss(shape = [5,5,3,64], stddev = 5e-2, wl = 0.0)
kernel1 = tf.nn.conv2d(image_holder, weight1, [1,1,1,1], padding = 'SAME')
bias1 = tf.Variable(tf.constant(0.0, shape = [64]))
conv1 = tf.nn.relu(tf.nn.bias_add(kernel1, bias1))
pool1 = tf.nn.max_pool(conv1, ksize = [1,3,3,1], strides = [1,2,2,1], padding = 'SAME')
norm1 = tf.nn.lrn(pool1, 4, bias = 1.0, alpha = 0.001/9.0, beta = 0.75)

weight2 = variable_with_weight_loss(shape = [5,5,64,64], stddev = 5e-2, wl = 0.0)
kernel2 = tf.nn.conv2d(norm1, weight2, [1,1,1,1], padding = 'SAME')
bias2 = tf.Variable(tf.constant(0.1, shape = [64]))
conv2 = tf.nn.relu(tf.nn.bias_add(kernel2, bias2))
norm2 = tf.nn.lrn(conv2, 4, bias = 1.0, alpha = 0.001/9.0, beta = 0.75)
pool2 = tf.nn.max_pool(norm2, ksize = [1,3,3,1], strides = [1,2,2,1], padding = 'SAME')

reshape = tf.reshape(pool2, [batch_size, -1])
#比如pool2总共有a×b×c个数，要reshape成batch_size行 × ???列，这里的-1就表示???列，tf.reshape会自行算出需要多少列a×b×c/batch_size
dim = reshape.get_shape()[1].value
weight3 = variable_with_weight_loss(shape = [dim, 384], stddev = 0.04, wl = 0.004)
bias3 = tf.Variable(tf.constant(0.1, shape = [384]))
local3 = tf.nn.relu(tf.matmul(reshape, weight3) + bias3)

weight4 = variable_with_weight_loss(shape = [384,192], stddev = 0.04, wl = 0.004)
bias4 = tf.Variable(tf.constant(0.1, shape = [192]))
local4 = tf.nn.relu(tf.matmul(local3, weight4) + bias4)

weight5 = variable_with_weight_loss(shape = [192,10], stddev = 1/192.0, wl = 0.0)
bias5 = tf.Variable(tf.constant(0.0, shape = [10]))
logits = tf.nn.relu(tf.matmul(local4, weight5) + bias5)

def loss(logits, labels):
	labels = tf.cast(labels, tf.int64)
	cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(logits = logits, labels = labels, name = 'cross_entropy_per_example')
	cross_entropy_mean = tf.reduce_mean(cross_entropy, name = 'cross_entropy')
	tf.add_to_collection('losses', cross_entropy_mean)
	return tf.add_n(tf.get_collection('losses'), name = 'total_loss')

loss = loss(logits, label_holder)
train_op = tf.train.AdamOptimizer(1e-3).minimize(loss)
top_k_op = tf.nn.in_top_k(logits, label_holder, 1)

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

tf.train.start_queue_runners()

for step in range(max_steps):
	start_time = time.time()
	image_batch, label_batch = sess.run([images_train, labels_train])
	_, loss_value = sess.run([train_op, loss], feed_dict = {image_holder: image_batch, label_holder: label_batch})
	duration = time.time() - start_time
	if step %10 ==0:
		examples_per_sec = batch_size / duration
		sec_per_batch = float(duration)

		format_str = ('step %d,loss=%.2f (%.1f example/sec; %.3f sec/batch)')
		print(format_str % (step, loss_value, examples_per_sec, sec_per_batch))

num_examples = 10000
import math
num_iter = int(math.ceil(num_examples / batch_size))
true_count = 0
total_sample_count = num_iter * batch_size
step = 0
while step < num_iter:
	image_batch, label_batch = sess.run([images_test,labels_test])
	predictions = sess.run([top_k_op], feed_dict = {image_holder: image_batch, label_holder: label_batch})
	true_count += np.sum(predictions)
	step += 1

precision = true_count / total_sample_count
print('precision @ 1 = %.3f' % precision)

max_step=100,num_example=100的运行结果（CPU版）

Filling queue with 20000 CIFAR images before starting to train. This will take a few minutes.
step 0,loss=4.67 (26.4 example/sec; 4.854 sec/batch)
step 10,loss=3.81 (127.4 example/sec; 1.005 sec/batch)
step 20,loss=3.20 (118.6 example/sec; 1.079 sec/batch)
step 30,loss=2.86 (128.0 example/sec; 1.000 sec/batch)
step 40,loss=2.59 (125.0 example/sec; 1.024 sec/batch)
step 50,loss=2.48 (125.8 example/sec; 1.018 sec/batch)
step 60,loss=2.34 (126.3 example/sec; 1.013 sec/batch)
step 70,loss=2.28 (128.4 example/sec; 0.997 sec/batch)
step 80,loss=2.30 (128.4 example/sec; 0.997 sec/batch)
step 90,loss=2.29 (127.8 example/sec; 1.002 sec/batch)
precision @ 1 = 0.297
[Finished in 108.8s]