Tensorflow+CIFAR-10实例讲解

本人使用的环境是：TensorFlow1.1.0+python3.6+GPU

CIFAR-10：由60000张32*32的RGB彩色图片构成，共10个分类。50000张训练，10000张测试（交叉验证）

模型：CNN

载入常用库，并载入TensorFlow Models中自动下载、读取CIFAR-10数据的类

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

定义batch_size、训练轮数max_steps，以及下载CIFAR-10数据的路径

max_steps = 3000
batch_size = 128
data_dir = '/tmp/cifar10_data/cifar-10-batches-bin'

定义初始化weight的函数，使用tf.truncated_normal截断的正态分布来初始化权重

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

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')

使用cifar10类下载数据集，并解压、展开到默认位置

cifar10.maybe_download_and_extract()

使用cifar10_input类中的distorted_inputs函数产生训练需要使用的数据，包括特征及其对应的label，每次执行都会生成一个batch_size的数量的样本

images_train, labels_train = cifar10_input.distorted_inputs(data_dir=data_dir,
                                                            batch_size=batch_size)

使用cifar10_input类中的inputs函数生成测试数据

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

创建输入数据的placeholder，包括特征和label

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

创建第一个卷积层，5*5的卷积核，3个颜色通道，64个卷积核，设置weight初始化函数的标准差为0.05；在ReLU之后，使用一个尺寸为3*3且步长为2*2的最大池化层

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)

创建第二个卷积层，bias值全部初始化为0.1，并调换最大池化层和LRN层的顺序

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')

第一个全连接层，对全连接层的weight进行初始化，隐含节点为384，正态分布的标准差为0.04，bias的值也初始化为0.1

reshape = tf.reshape(pool2, [batch_size, -1])
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)

第二个全连接层，其隐含节点数下降一半，为192个，其他超参数保持不变

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)

softmax层

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.add(tf.matmul(local4, weight5), bias5)

至此完成了整个网络inference的部分

接着将logits节点和label_placeholder传入loss函数获得最终的loss

loss = loss(logits, label_holder)

优化器选择Adam Optimizer，学习速率设为le-3

train_op = tf.train.AdamOptimizer(1e-3).minimize(loss)

求输出结果中top k的准确率，默认使用top 1

top_k_op = tf.nn.in_top_k(logits, label_holder, 1)

创建默认的session，并初始化全部模型参数

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

启动图片数据增强的线程队列（16个）

tf.train.start_queue_runners()

开始训练，首先使用session的run方法执行image_train、label_train的计算，获得一个batch的训练数据，再将这个batch的数据传入train_op和loss的计算

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 examples/sec; %.3f sec/batch)')
        print(format_str % (step, loss_value, examples_per_sec, sec_per_batch))

评测模型在测试集上的准确率，测试集一共有10000个样本

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

计算准确率的评测结果并打印

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

本人运行结果如下

另外，将迭代数修改为30000时，得到结果如下

注：缺失的Python函数可取github网站上搜索下载。

此处列出一个地址，参考使用

https://github.com/tensorflow/models/tree/v1.10.0/tutorials/image/cifar10