tensorflow cifar10 代码详解

刚看时看代码比价懵逼，还好其实不难。写出来注释加深理解一下。
代码主要分为4个py文件。详细请见代码的注释。
注：由于tensorflow进行了更新，一些函数名已经在代码中改动。
input.py

 # -*- coding: utf-8 -*- 
# Copyright 2015 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

"""Routine for decoding the CIFAR-10 binary file format."""

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import os

import tensorflow.python.platform
from six.moves import xrange  # pylint: disable=redefined-builtin
import tensorflow as tf

from tensorflow.python.platform import gfile

# Process images of this size. Note that this differs from the original CIFAR
# image size of 32 x 32. If one alters 改变 this number, then the entire model
# architecture will change and any model would need to be retrained.
IMAGE_SIZE = 24

# Global constants describing the CIFAR-10 data set.

NUM_CLASSES = 10
NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN = 50000
NUM_EXAMPLES_PER_EPOCH_FOR_EVAL = 10000


def read_cifar10(filename_queue):
  """Reads and parses 从语法上分析或描述 examples from CIFAR10 data files.

  Recommendation: if you want N-way read parallelism, call this function
  N times.  This will give you N independent Readers reading different
  files & positions within those files, which will give better mixing 混合 of
  examples.

  Args:
    filename_queue: A queue of strings with the filenames to read from.

  Returns:
    An object representing a single example, with the following fields:
    默认定义 行是高，列是宽，通道是深度。
      height: number of rows in the result (32)
      width: number of columns in the result (32)
      depth: number of color channels in the result (3)
      key: a scalar 标量的 string Tensor describing the filename & record number
        for this example.
      label: an int32 Tensor with the label in the range 0..9.
      uint8image: a [height, width, depth] uint8 Tensor with the image data
  """

  class CIFAR10Record(object):
    pass
  # 这个result 是什么意思。。。   定义一个空类，类似于C里面的结构体。
  result = CIFAR10Record()

  # Dimensions of the images in the CIFAR-10 dataset.
  # See http://www.cs.toronto.edu/~kriz/cifar.html for a description of the
  # input format.
  label_bytes = 1  # 2 for CIFAR-100
  result.height = 32
  result.width = 32
  result.depth = 3
  image_bytes = result.height * result.width * result.depth
  # Every record consists of a label followed by the image, with a
  # fixed number of bytes for each.
  record_bytes = label_bytes + image_bytes

  # Read a record, getting filenames from the filename_queue.  No
  # header or footer 页脚 in the CIFAR-10 format, so we leave header_bytes
  # and footer_bytes at their default of 0.
  #读取固定长度的函数  好像是二进制文件。
  reader = tf.FixedLengthRecordReader(record_bytes=record_bytes)
  #返回从filename_queue 得到的（key，value）对。key是value 是字符串型的tensor。
  # 所以这个key 到底是干嘛的？
  result.key, value = reader.read(filename_queue)

  # Convert from a string to a vector of uint8 that is record_bytes long.
  #在这里直接改变了record_bytes 的类型。个人认为这样不太好。应该是可以重新来一个新的变量吧？
  record_bytes = tf.decode_raw(value, tf.uint8)

  # The first bytes represent the label, which we convert from uint8->int32.
  #tf.cast 是 拼接， tf.slice是切片。0代表了起点。lable_bytes 代表了长度。 但是为什么长度为1 ？？
  #因为这个tf.slice是从一维的tensor中截取一个slice。而record_bytes 现在是很多个一维向量组成的。
  result.label = tf.cast(
      tf.slice(record_bytes, [0], [label_bytes]), tf.int32)

  # The remaining bytes after the label represent the image, which we reshape
  # from [depth * height * width] to [depth, height, width].
  # 这个lable_bytes 到底是数，还是向量。。。。。？？ 看错了。前面变的是record_bytes。。。
  depth_major = tf.reshape(tf.slice(record_bytes, [label_bytes], [image_bytes]),
                           [result.depth, result.height, result.width])
  # Convert from [depth, height, width] to [height, width, depth].
  result.uint8image = tf.transpose(depth_major, [1, 2, 0])

  return result

#目前为止 这个函数是 正确的得到了 lable 和 image 。
#http://blog.csdn.net/diligent_321/article/details/53130913
def _generate_image_and_label_batch(image, label, min_queue_examples,
                                    batch_size):
  """Construct a queued batch of images and labels.

  Args:
    image: 3-D Tensor of [height, width, 3] of type.float32.
    label: 1-D Tensor of type.int32
    min_queue_examples: int32, minimum number of samples to retain
      in the queue that provides of batches of examples.
    batch_size: Number of images per batch.

  Returns:
    images: Images. 4D tensor of [batch_size, height, width, 3] size.
    #lables 是不是不需要更改。？
    labels: Labels. 1D tensor of [batch_size] size.
  """
  # Create a queue that shuffles the examples, and then
  # read 'batch_size' images + labels from the example queue.
  num_preprocess_threads = 16
  images, label_batch = tf.train.shuffle_batch(
      [image, label], # 这里为什么加一个lable 应该是为了生成 lable_batch. ？
      batch_size=batch_size,
      num_threads=num_preprocess_threads,
      capacity=min_queue_examples + 3 * batch_size,
      min_after_dequeue=min_queue_examples)

  # Display the training images in the visualizer 观察仪.
  tf.summary.image('images', images)
  # 现在返回的是 images 和 lable_batch 也就是打乱后的结果。
  return images, tf.reshape(label_batch, [batch_size])


def distorted_inputs(data_dir, batch_size):
  """Construct distorted 扭曲的 input for CIFAR training using the Reader ops.

  Args:
    data_dir: Path to the CIFAR-10 data directory.
    batch_size: Number of images per batch.

  Returns:
    images: Images. 4D tensor of [batch_size, IMAGE_SIZE, IMAGE_SIZE, 3] size.
    labels: Labels. 1D tensor of [batch_size] size.
  """
  filenames = [os.path.join(data_dir, 'data_batch_%d.bin' % i)
               for i in xrange(1, 6)]
  for f in filenames:
    if not gfile.Exists(f):
      raise ValueError('Failed to find file: ' + f)

  # Create a queue that produces the filenames to read.
  filename_queue = tf.train.string_input_producer(filenames)

  # Read examples from files in the filename queue.
  read_input = read_cifar10(filename_queue)
  # 刚刚说错了，tf.cast是转化格式的函数。
  reshaped_image = tf.cast(read_input.uint8image, tf.float32)

  height = IMAGE_SIZE
  width = IMAGE_SIZE

  # Image processing for training the network. Note the many random
  # distortions applied to the image.
  #为什么是随机的生成一个部分？
  # Randomly crop a [height, width] section of the image.
  #distorted_image = tf.image.random_crop(reshaped_image, [height, width])
  distorted_image = tf.random_crop(reshaped_image, [height, width, 3])

  # Randomly flip 弹 筋斗 the image horizontally 水平的.
  #图片随机左右翻转。
  distorted_image = tf.image.random_flip_left_right(distorted_image)

  # Because these operations are not commutative 交换的，交替的, consider randomizing
  # randomize the order their operation.
  #随机的调节亮度
  distorted_image = tf.image.random_brightness(distorted_image,
                                               max_delta=63)
  #随机调节对比度。 contrast 对比。
  distorted_image = tf.image.random_contrast(distorted_image,
                                             lower=0.2, upper=1.8)

  # Subtract off the mean and divide by the variance of the pixels.

  #float_image = tf.image.per_image_whitening(distorted_image)
  float_image = tf.image.per_image_standardization(distorted_image)
  # Ensure that the random shuffling has good mixing properties.
  min_fraction_of_examples_in_queue = 0.4
  min_queue_examples = int(NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN *
                           min_fraction_of_examples_in_queue)
  print ('Filling queue with %d CIFAR images before starting to train. '
         'This will take a few minutes.' % min_queue_examples)

  # Generate a batch of images and labels by building up a queue of examples.
  return _generate_image_and_label_batch(float_image, read_input.label,
                                         min_queue_examples, batch_size)


def inputs(eval_data, data_dir, batch_size):
  """Construct input for CIFAR evaluation using the Reader ops.

  Args:
    eval_data: bool, indicating if one should use the train or eval data set.
    data_dir: Path to the CIFAR-10 data directory.
    batch_size: Number of images per batch.

  Returns:
    images: Images. 4D tensor of [batch_size, IMAGE_SIZE, IMAGE_SIZE, 3] size.
    labels: Labels. 1D tensor of [batch_size] size.
  """
  if not eval_data:
    filenames = [os.path.join(data_dir, 'data_batch_%d.bin' % i)
                 for i in xrange(1, 6)]
    num_examples_per_epoch = NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN
  else:
    filenames = [os.path.join(data_dir, 'test_batch.bin')]
    num_examples_per_epoch = NUM_EXAMPLES_PER_EPOCH_FOR_EVAL

  for f in filenames:
    if not gfile.Exists(f):
      raise ValueError('Failed to find file: ' + f)

  # Create a queue that produces the filenames to read.
  filename_queue = tf.train.string_input_producer(filenames)

  # Read examples from files in the filename queue.
  read_input = read_cifar10(filename_queue)
  reshaped_image = tf.cast(read_input.uint8image, tf.float32)

  height = IMAGE_SIZE
  width = IMAGE_SIZE

  # Image processing for evaluation.
  # Crop the central [height, width] of the image.
  resized_image = tf.image.resize_image_with_crop_or_pad(reshaped_image,
                                                         width, height)

  # Subtract off the mean and divide by the variance of the pixels.
  float_image = tf.image.per_image_standardization(resized_image)

  # Ensure that the random shuffling has good mixing properties.
  min_fraction_of_examples_in_queue = 0.4
  min_queue_examples = int(num_examples_per_epoch *
                           min_fraction_of_examples_in_queue)

  # Generate a batch of images and labels by building up a queue of examples.
  return _generate_image_and_label_batch(float_image, read_input.label,
                                         min_queue_examples, batch_size)

cifar10.py

 # -*- coding: utf-8 -*- 
# Copyright 2015 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

"""Builds the CIFAR-10 network.

Summary of available functions:

 # Compute input images and labels for training. If you would like to run
 # evaluations, use input() instead.
 inputs, labels = distorted_inputs()

 # Compute inference 预测 on the model inputs to make a prediction.
 predictions = inference(inputs)

 # Compute the total loss of the prediction with respect to the labels.
 loss = loss(predictions, labels)

 # Create a graph to run one step of training with respect to the loss.
 train_op = train(loss, global_step)
"""
# pylint: disable=missing-docstring
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import gzip
import os
import re
import sys
import tarfile

import tensorflow.python.platform
from six.moves import urllib
import tensorflow as tf

#from tensorflow.models.image.cifar10 import cifar10_input
import cifar10_input

FLAGS = tf.app.flags.FLAGS

# Basic model parameters 参数 .
tf.app.flags.DEFINE_integer('batch_size', 128,
                            """Number of images to process in a batch.""")
tf.app.flags.DEFINE_string('data_dir', 'tmp/cifar10_data',
                           """Path to the CIFAR-10 data directory.""")

# Global constants describing the CIFAR-10 data set.
IMAGE_SIZE = cifar10_input.IMAGE_SIZE #应该是24。
NUM_CLASSES = cifar10_input.NUM_CLASSES # 10 个类
NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN = cifar10_input.NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN
NUM_EXAMPLES_PER_EPOCH_FOR_EVAL = cifar10_input.NUM_EXAMPLES_PER_EPOCH_FOR_EVAL


# Constants 常数 describing the training process.
MOVING_AVERAGE_DECAY = 0.9999     # The decay 腐烂 衰退 to use for the moving average.
NUM_EPOCHS_PER_DECAY = 350.0      # Epochs after which learning rate decays. 阶梯宽度？？
LEARNING_RATE_DECAY_FACTOR = 0.1  # Learning rate decay factor.
INITIAL_LEARNING_RATE = 0.1       # Initial learning rate.

# If a model is trained with multiple GPU's prefix 前缀 all Op names with tower_name
# to differentiate the operations. Note that this prefix is removed from the
# names of the summaries when visualizing a model.
TOWER_NAME = 'tower'

DATA_URL = 'http://www.cs.toronto.edu/~kriz/cifar-10-binary.tar.gz'

# 什么意思？？？？？？？？？？？？？
def _activation_summary(x):
  """Helper to create summaries for activations.

  Creates a summary that provides a histogram 直方图 of activations 激活 .
  Creates a summary that measure the sparsity 稀疏 of activations.

  Args:
    x: Tensor
  Returns:
    nothing
  """
  # Remove 'tower_[0-9]/' from the name in case this is a multi-GPU training
  # session. This helps the clarity of presentation on tensorboard.
  tensor_name = re.sub('%s_[0-9]*/' % TOWER_NAME, '', x.op.name)
  # 第一个是 histogram 是记录变量， scalar 是记录标量。
  tf.summary.histogram(tensor_name + '/activations', x)
  tf.summary.scalar(tensor_name + '/sparsity', tf.nn.zero_fraction(x))


def _variable_on_cpu(name, shape, initializer):
  """Helper to create a Variable stored on CPU memory.

  Args:
    name: name of the variable
    shape: list of ints
    initializer: initializer for Variable

  Returns:
    Variable Tensor
  """
  with tf.device('/cpu:0'):
    var = tf.get_variable(name, shape, initializer=initializer)
  return var


def _variable_with_weight_decay(name, shape, stddev, wd):
  """Helper to create an initialized Variable with weight decay.

  Note that the Variable is initialized with a truncated 切去顶端的 缩短了的 normal distribution.
  A weight decay is added only if one is specified.

  Args:
    name: name of the variable
    shape: list of ints
    stddev: standard deviation of a truncated Gaussian
    wd: add L2Loss weight decay multiplied by this float. If None, weight
        decay is not added for this Variable.

  Returns:
    Variable Tensor
  """
  # 从截断的正态分布中输出随机值。 
  var = _variable_on_cpu(name, shape,
                         tf.truncated_normal_initializer(stddev=stddev))
  if wd:
    # 进行l2 正则化。
    weight_decay = tf.multiply(tf.nn.l2_loss(var), wd, name='weight_loss')
    # 加入到loss里。 ？？
    tf.add_to_collection('losses', weight_decay)
  return var


def distorted_inputs():
  """Construct distorted input for CIFAR training using the Reader ops.

  Returns:
    images: Images. 4D tensor of [batch_size, IMAGE_SIZE, IMAGE_SIZE, 3] size.
    labels: Labels. 1D tensor of [batch_size] size.

  Raises:
    ValueError: If no data_dir
  """
  if not FLAGS.data_dir:
    raise ValueError('Please supply a data_dir')
  data_dir = os.path.join(FLAGS.data_dir, 'cifar-10-batches-bin')
  return cifar10_input.distorted_inputs(data_dir=data_dir,
                                        batch_size=FLAGS.batch_size)

# 怎么感觉这个和cifar10_input 有那么重复定义的感觉。 ？
def inputs(eval_data):
  """Construct input for CIFAR evaluation using the Reader ops.

  Args:
    eval_data: bool, indicating if one should use the train or eval data set.

  Returns:
    images: Images. 4D tensor of [batch_size, IMAGE_SIZE, IMAGE_SIZE, 3] size.
    labels: Labels. 1D tensor of [batch_size] size.

  Raises:
    ValueError: If no data_dir
  """
  if not FLAGS.data_dir:
    raise ValueError('Please supply a data_dir')
  data_dir = os.path.join(FLAGS.data_dir, 'cifar-10-batches-bin')
  return cifar10_input.inputs(eval_data=eval_data, data_dir=data_dir,
                              batch_size=FLAGS.batch_size)


def inference(images):
  """Build the CIFAR-10 model.

  Args:
    4D-tensor
    images: Images returned from distorted_inputs() or inputs().

  Returns:
    Logits.
  """
  # We instantiate all variables using tf.get_variable() instead of
  # tf.Variable() in order to share variables across multiple GPU training runs.
  # If we only ran this model on a single GPU, we could simplify this function
  # by replacing all instances of tf.get_variable() with tf.Variable().
  #
  # conv1
  with tf.variable_scope('conv1') as scope:
    #生成  卷积核。
    kernel = _variable_with_weight_decay('weights', shape=[5, 5, 3, 64],
                                         stddev=1e-4, wd=0.0)
    #开始卷积。
    conv = tf.nn.conv2d(images, kernel, [1, 1, 1, 1], padding='SAME')
    biases = _variable_on_cpu('biases', [64], tf.constant_initializer(0.0))
    bias = tf.nn.bias_add(conv, biases)
    conv1 = tf.nn.relu(bias, name=scope.name)
    _activation_summary(conv1)

  # pool1
  #第一个参数是feature map， 第二个ksize 池化窗口的大小为[1,heigt,width,1],
  #第三个是 [1, stride,stride, 1]， padding 决定是不是 pooling 后size是不是变。
  pool1 = tf.nn.max_pool(conv1, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1],
                         padding='SAME', name='pool1')
  # norm1
  #tf.nn.lrn(input,depth_radius=None,bias=None,alpha=None,beta=None,name=None)
  # 局部响应归一化。。 回去再看具体内容。
  norm1 = tf.nn.lrn(pool1, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75,
                    name='norm1')

  # conv2
  with tf.variable_scope('conv2') as scope:
    #依然是64通道。
    kernel = _variable_with_weight_decay('weights', shape=[5, 5, 64, 64],
                                         stddev=1e-4, wd=0.0)
    conv = tf.nn.conv2d(norm1, kernel, [1, 1, 1, 1], padding='SAME')
    biases = _variable_on_cpu('biases', [64], tf.constant_initializer(0.1))
    #将偏差加到value上。
    bias = tf.nn.bias_add(conv, biases)
    conv2 = tf.nn.relu(bias, name=scope.name)
    _activation_summary(conv2)

  # norm2
  norm2 = tf.nn.lrn(conv2, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75,
                    name='norm2')
  # pool2
  pool2 = tf.nn.max_pool(norm2, ksize=[1, 3, 3, 1],
                         strides=[1, 2, 2, 1], padding='SAME', name='pool2')

  # local3
  with tf.variable_scope('local3') as scope:
    # Move everything into depth so we can perform a single matrix multiply.
    dim = 1
    for d in pool2.get_shape()[1:].as_list():
      dim *= d
    reshape = tf.reshape(pool2, [FLAGS.batch_size, dim])
    #为什么是 384??
    weights = _variable_with_weight_decay('weights', shape=[dim, 384],
                                          stddev=0.04, wd=0.004)
    biases = _variable_on_cpu('biases', [384], tf.constant_initializer(0.1))
    local3 = tf.nn.relu(tf.matmul(reshape, weights) + biases, name=scope.name)
    _activation_summary(local3)

  # local4
  with tf.variable_scope('local4') as scope:
    weights = _variable_with_weight_decay('weights', shape=[384, 192],
                                          stddev=0.04, wd=0.004)
    biases = _variable_on_cpu('biases', [192], tf.constant_initializer(0.1))
    local4 = tf.nn.relu(tf.matmul(local3, weights) + biases, name=scope.name)
    _activation_summary(local4)

  # softmax, i.e. softmax(WX + b)
  # 不是很清楚为什么要192。  ？
  with tf.variable_scope('softmax_linear') as scope:
    weights = _variable_with_weight_decay('weights', [192, NUM_CLASSES],
                                          stddev=1/192.0, wd=0.0)
    biases = _variable_on_cpu('biases', [NUM_CLASSES],
                              tf.constant_initializer(0.0))
    softmax_linear = tf.add(tf.matmul(local4, weights), biases, name=scope.name)
    _activation_summary(softmax_linear)

  return softmax_linear


def loss(logits, labels):
  """Add L2Loss to all the trainable variables.

  Add summary for for "Loss" and "Loss/avg".
  Args:
    logits: Logits from inference().
    labels: Labels from distorted_inputs or inputs(). 1-D tensor
            of shape [batch_size]

  Returns:
    Loss tensor of type float.
  """
  # Reshape the labels into a dense 稠密的 Tensor of
  # shape [batch_size, NUM_CLASSES]. sparse 稀疏的。
  #把 行向量 变成 列向量。
  sparse_labels = tf.reshape(labels, [FLAGS.batch_size, 1])
  # 把行向量 reshape 成列向量。
  indices = tf.reshape(tf.range(FLAGS.batch_size), [FLAGS.batch_size, 1])
  #在第二个维度上链接。 也就是在列上链接。
  #concated = tf.concat(1, [indices, sparse_labels])
  concated = tf.concat([indices, sparse_labels],1)
  # 怎么好像是 dense to spare 。
  dense_labels = tf.sparse_to_dense(concated,
                                    [FLAGS.batch_size, NUM_CLASSES],
                                    1.0, 0.0)

  # Calculate the average cross entropy loss across the batch.
  cross_entropy = tf.nn.softmax_cross_entropy_with_logits(
      logits = logits, labels = dense_labels, name='cross_entropy_per_example')
  cross_entropy_mean = tf.reduce_mean(cross_entropy, name='cross_entropy')
  #tf.add_to_collection：把变量放入一个集合，把很多变量变成一个列表 #tf.get_collection：从一个结合中取出全部变量，是一个列表
  tf.add_to_collection('losses', cross_entropy_mean)

  # The total loss is defined as the cross entropy loss plus all of the weight
  # decay terms (L2 loss).
  #tf.add_n([p1, p2, p3....])函数是实现一个列表，矩阵 的元素的相加
  #把所有的元素取出来，然后加起来。
  return tf.add_n(tf.get_collection('losses'), name='total_loss')


def _add_loss_summaries(total_loss):
  """Add summaries for losses in CIFAR-10 model.

  Generates moving average for all losses and associated summaries for
  visualizing the performance of the network.

  Args:
    total_loss: Total loss from loss().
  Returns:
    loss_averages_op: op for generating moving averages of losses.
  """
  # Compute the moving average of all individual losses and the total loss.
  #创建一个新的指数滑动均值对象
  loss_averages = tf.train.ExponentialMovingAverage(0.9, name='avg')
  # 从字典集合中返回关键字'losses'对应的所有变量，包括交叉熵损失和正则项损失  
  losses = tf.get_collection('losses')
  # 这一步 ？ # 创建‘shadow variables’,并添加维护滑动均值的操作  
  loss_averages_op = loss_averages.apply(losses + [total_loss])

  # Attach a scalar summary to all individual losses and the total loss; do the
  # same for the averaged version of the losses.
  for l in losses + [total_loss]:
    # Name each loss as '(raw)' and name the moving average version of the loss
    # as the original loss name.
    # ？
    tf.summary.scalar(l.op.name +' (raw)', l)
    tf.summary.scalar(l.op.name, loss_averages.average(l))

  return loss_averages_op


def train(total_loss, global_step):
  """Train CIFAR-10 model.

  Create an optimizer and apply to all trainable variables. Add moving
  average for all trainable variables.

  Args:
    total_loss: Total loss from loss().
    global_step: Integer Variable counting the number of training steps
      processed.
  Returns:
    train_op: op for training.
  """
  # Variables that affect learning rate.
  num_batches_per_epoch = NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN / FLAGS.batch_size
  #每经过decay_step步训练，衰减lr  
  decay_steps = int(num_batches_per_epoch * NUM_EPOCHS_PER_DECAY)

  # Decay the learning rate exponentially based on the number of steps.
  lr = tf.train.exponential_decay(INITIAL_LEARNING_RATE,
                                  global_step,
                                  decay_steps,
                                  LEARNING_RATE_DECAY_FACTOR,
                                  staircase=True)
  tf.summary.scalar('learning_rate', lr)

  # Generate moving averages of all losses and associated summaries.
  loss_averages_op = _add_loss_summaries(total_loss)

  # Compute gradients.
  #tf.control_dependencies是一个context manager,控制节点执行顺序，先执行control_inputs中的操作，再执行context中的操作 
  with tf.control_dependencies([loss_averages_op]):
    opt = tf.train.GradientDescentOptimizer(lr)
    #返回计算出的（gradient, variable） pairs  ？
    grads = opt.compute_gradients(total_loss)

  # Apply gradients.
  #返回一步梯度更新操作   ？
  apply_gradient_op = opt.apply_gradients(grads, global_step=global_step)

  # Add histograms for trainable variables.
  for var in tf.trainable_variables():
    tf.summary.histogram(var.op.name, var)

  # Add histograms for gradients.
  for grad, var in grads:
    if grad is not None:
      tf.summary.histogram(var.op.name + '/gradients', grad)

  # Track the moving averages of all trainable variables.
  variable_averages = tf.train.ExponentialMovingAverage(
      MOVING_AVERAGE_DECAY, global_step)
  # tf.trainable_variables() 方法返回所有`trainable=True`的变量，列表结构 
  #返回模型参数变量的滑动更新操作  
  variables_averages_op = variable_averages.apply(tf.trainable_variables())

  with tf.control_dependencies([apply_gradient_op, variables_averages_op]):
    train_op = tf.no_op(name='train')
    #Does nothing. Only useful as a placeholder for control edges

  return train_op


def maybe_download_and_extract():
  """Download and extract the tarball from Alex's website."""
  dest_directory = FLAGS.data_dir
  if not os.path.exists(dest_directory):
    os.makedirs(dest_directory)
  filename = DATA_URL.split('/')[-1]
  filepath = os.path.join(dest_directory, filename)
  if not os.path.exists(filepath):
    def _progress(count, block_size, total_size):
      sys.stdout.write('\r>> Downloading %s %.1f%%' % (filename,
          float(count * block_size) / float(total_size) * 100.0))
      sys.stdout.flush()
    filepath, _ = urllib.request.urlretrieve(DATA_URL, filepath,
                                             reporthook=_progress)
    print()
    statinfo = os.stat(filepath)
    print('Successfully downloaded', filename, statinfo.st_size, 'bytes.')
    tarfile.open(filepath, 'r:gz').extractall(dest_directory)

train.py

 # -*- coding: utf-8 -*- 
# Copyright 2015 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

"""A binary to train CIFAR-10 using a single GPU.

Accuracy:
cifar10_train.py achieves ~86% accuracy after 100K steps (256 epochs of
data) as judged by cifar10_eval.py.

Speed: With batch_size 128.

System        | Step Time (sec/batch)  |     Accuracy
------------------------------------------------------------------
1 Tesla K20m  | 0.35-0.60              | ~86% at 60K steps  (5 hours)
1 Tesla K40m  | 0.25-0.35              | ~86% at 100K steps (4 hours)

Usage:
Please see the tutorial and website for how to download the CIFAR-10
data set, compile the program and train the model.

http://tensorflow.org/tutorials/deep_cnn/
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from datetime import datetime
import os.path
import time

import tensorflow.python.platform
from tensorflow.python.platform import gfile

import numpy as np
from six.moves import xrange  # pylint: disable=redefined-builtin
import tensorflow as tf

#from tensorflow.models.image.cifar10 import cifar10
import cifar10

FLAGS = tf.app.flags.FLAGS

tf.app.flags.DEFINE_string('train_dir', 'tmp/cifar10_train',
                           """Directory where to write event logs """
                           """and checkpoint.""")
tf.app.flags.DEFINE_integer('max_steps', 2000,
                            """Number of batches to run.""")
tf.app.flags.DEFINE_boolean('log_device_placement', False,
                            """Whether to log device placement.""")


def train():
  """Train CIFAR-10 for a number of steps."""
  with tf.Graph().as_default():
    global_step = tf.Variable(0, trainable=False)

    # Get images and labels for CIFAR-10.
    images, labels = cifar10.distorted_inputs()

    # Build a Graph that computes the logits predictions from the
    # inference model.
    logits = cifar10.inference(images)

    # Calculate loss.
    loss = cifar10.loss(logits, labels)

    # Build a Graph that trains the model with one batch of examples and
    # updates the model parameters.
    train_op = cifar10.train(loss, global_step)

    # Create a saver.
    saver = tf.train.Saver(tf.all_variables())

    # Build the summary operation based on the TF collection of Summaries.
    summary_op = tf.summary.merge_all()

    # Build an initialization operation to run below.
    init = tf.initialize_all_variables()

    # Start running operations on the Graph.
    #log_device_placement参数可以记录每一个操作使用的设备，这里的操作比较多，就不需要记录了，故设置为False  
    sess = tf.Session(config=tf.ConfigProto(
        log_device_placement=FLAGS.log_device_placement))
    sess.run(init)

    # Start the queue runners.
    tf.train.start_queue_runners(sess=sess)

    summary_writer = tf.summary.FileWriter(FLAGS.train_dir,
                                            graph_def=sess.graph_def)

    for step in xrange(FLAGS.max_steps):
      start_time = time.time()
      _, loss_value = sess.run([train_op, loss])
      duration = time.time() - start_time
       #用于验证当前迭代计算出的loss_value是否合理  
      assert not np.isnan(loss_value), 'Model diverged with loss = NaN'

      if step % 10 == 0:
        num_examples_per_step = FLAGS.batch_size
        examples_per_sec = num_examples_per_step / duration
        sec_per_batch = float(duration)

        format_str = ('%s: step %d, loss = %.2f (%.1f examples/sec; %.3f '
                      'sec/batch)')
        print (format_str % (datetime.now(), step, loss_value,
                             examples_per_sec, sec_per_batch))

      if step % 100 == 0:
        summary_str = sess.run(summary_op)
        summary_writer.add_summary(summary_str, step)

      # Save the model checkpoint periodically.
      if step % 1000 == 0 or (step + 1) == FLAGS.max_steps:
        checkpoint_path = os.path.join(FLAGS.train_dir, 'model.ckpt')
        saver.save(sess, checkpoint_path, global_step=step)


def main(argv=None):  # pylint: disable=unused-argument
  cifar10.maybe_download_and_extract()
  if gfile.Exists(FLAGS.train_dir):
    gfile.DeleteRecursively(FLAGS.train_dir)
  gfile.MakeDirs(FLAGS.train_dir)
  train()


if __name__ == '__main__':
  tf.app.run()

eval.py

 # -*- coding: utf-8 -*- 
# Copyright 2015 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

"""Evaluation for CIFAR-10.

Accuracy:
cifar10_train.py achieves 83.0% accuracy after 100K steps (256 epochs
of data) as judged by cifar10_eval.py.

Speed:
On a single Tesla K40, cifar10_train.py processes a single batch of 128 images
in 0.25-0.35 sec (i.e. 350 - 600 images /sec). The model reaches ~86%
accuracy after 100K steps in 8 hours of training time.

Usage:
Please see the tutorial and website for how to download the CIFAR-10
data set, compile the program and train the model.

http://tensorflow.org/tutorials/deep_cnn/
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from datetime import datetime
import math
import time

import tensorflow.python.platform
from tensorflow.python.platform import gfile
import numpy as np
import tensorflow as tf

#from tensorflow.models.image.cifar10 import cifar10
import cifar10

FLAGS = tf.app.flags.FLAGS

tf.app.flags.DEFINE_string('eval_dir', 'tmp/cifar10_eval',
                           """Directory where to write event logs.""")
tf.app.flags.DEFINE_string('eval_data', 'test',
                           """Either 'test' or 'train_eval'.""")
tf.app.flags.DEFINE_string('checkpoint_dir', 'tmp/cifar10_train',
                           """Directory where to read model checkpoints.""")
tf.app.flags.DEFINE_integer('eval_interval_secs', 60 * 5,
                            """How often to run the eval.""")
tf.app.flags.DEFINE_integer('num_examples', 10000,
                            """Number of examples to run.""")
tf.app.flags.DEFINE_boolean('run_once', False,
                         """Whether to run eval only once.""")


def eval_once(saver, summary_writer, top_k_op, summary_op):
  """Run Eval once.

  Args:
    saver: Saver.
    summary_writer: Summary writer.
    top_k_op: Top K op.
    summary_op: Summary op.
  """
  with tf.Session() as sess:
    ckpt = tf.train.get_checkpoint_state(FLAGS.checkpoint_dir)
    if ckpt and ckpt.model_checkpoint_path:
      # Restores from checkpoint
      saver.restore(sess, ckpt.model_checkpoint_path)
      # Assuming model_checkpoint_path looks something like:
      #   /my-favorite-path/cifar10_train/model.ckpt-0,
      # extract global_step from it.
      global_step = ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1]
    else:
      print('No checkpoint file found')
      return

    # Start the queue runners.
    coord = tf.train.Coordinator()
    try:
      threads = []
      for qr in tf.get_collection(tf.GraphKeys.QUEUE_RUNNERS):
        threads.extend(qr.create_threads(sess, coord=coord, daemon=True,
                                         start=True))

      num_iter = int(math.ceil(FLAGS.num_examples / FLAGS.batch_size))
      true_count = 0  # Counts the number of correct predictions.
      total_sample_count = num_iter * FLAGS.batch_size
      step = 0
      while step < num_iter and not coord.should_stop():
        predictions = sess.run([top_k_op])
        true_count += np.sum(predictions)
        step += 1

      # Compute precision @ 1.
      precision = true_count / total_sample_count
      print('%s: precision @ 1 = %.3f' % (datetime.now(), precision))

      summary = tf.Summary()
      summary.ParseFromString(sess.run(summary_op))
      summary.value.add(tag='Precision @ 1', simple_value=precision)
      summary_writer.add_summary(summary, global_step)
    except Exception as e:  # pylint: disable=broad-except
      coord.request_stop(e)

    coord.request_stop()
    coord.join(threads, stop_grace_period_secs=10)


def evaluate():
  """Eval CIFAR-10 for a number of steps."""
  with tf.Graph().as_default():
    # Get images and labels for CIFAR-10.
    eval_data = FLAGS.eval_data == 'test'
    images, labels = cifar10.inputs(eval_data=eval_data)

    # Build a Graph that computes the logits predictions from the
    # inference model.
    logits = cifar10.inference(images)

    # Calculate predictions.
    top_k_op = tf.nn.in_top_k(logits, labels, 1)

    # Restore the moving average version of the learned variables for eval.
    variable_averages = tf.train.ExponentialMovingAverage(
        cifar10.MOVING_AVERAGE_DECAY)
    variables_to_restore = variable_averages.variables_to_restore()
    saver = tf.train.Saver(variables_to_restore)

    # Build the summary operation based on the TF collection of Summaries.
    summary_op = tf.summary.merge_all()

    graph_def = tf.get_default_graph().as_graph_def()
    summary_writer = tf.summary.FileWriter(FLAGS.eval_dir,
                                            graph_def=graph_def)

    while True:
      eval_once(saver, summary_writer, top_k_op, summary_op)
      if FLAGS.run_once:
        break
      time.sleep(FLAGS.eval_interval_secs)


def main(argv=None):  # pylint: disable=unused-argument
  cifar10.maybe_download_and_extract()
  if gfile.Exists(FLAGS.eval_dir):
    gfile.DeleteRecursively(FLAGS.eval_dir)
  gfile.MakeDirs(FLAGS.eval_dir)
  evaluate()


if __name__ == '__main__':
  tf.app.run()