CIFAR10 代码分析详解——cifar10.py

引入库，定义各种参数

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

import os
import re
import sys
import tarfile

from six.moves import urllib
import tensorflow as tf

import cifar10_input

FLAGS = tf.app.flags.FLAGS

# Basic model parameters.
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.""")
tf.app.flags.DEFINE_boolean('use_fp16', False,
"""Train the model using fp16.""")

# Global constants describing the CIFAR-10 data set.
IMAGE_SIZE = cifar10_input.IMAGE_SIZE
NUM_CLASSES = cifar10_input.NUM_CLASSES
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 describing the training process.
MOVING_AVERAGE_DECAY = 0.9999     # The decay to use for the moving average.
NUM_EPOCHS_PER_DECAY = 350.0      # Epochs after which learning rate decays.
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 GPUs, prefix 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'

创建一个 summary 函数来记录 histogram 和 scalar，注意 scalar 记录的是x中0占的比例，用以衡量x的稀疏性sparsity。

def _activation_summary(x):
"""Helper to create summaries for activations.
Creates a summary that provides a histogram of activations.
Creates a summary that measures the sparsity 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)
tf.summary.histogram(tensor_name + '/activations', x)
tf.summary.scalar(tensor_name + '/sparsity',
tf.nn.zero_fraction(x))

创建 variable 生成函数，该变量存储在 cpu:0 上。输入参数 initializer 是用来指定何种方式初始化变量，比如：initializer=tf.constant_initializer(0.0)/tf.truncated_normal_initializer(stddev,dtype))。

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'):
dtype = tf.float16 if FLAGS.use_fp16 else tf.float32
var = tf.get_variable(name, shape, initializer=initializer, dtype=dtype)
return var

创建 variable 生成函数，该函数与上面的函数不同，只用以生成制定标准差的正态分布变量，此外，如果想对该变量进行 weight decay，需要指定参数 wd，并把 weight decay 项用

tf.add_to_collection 加入到 'losses' 中。

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 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
"""
dtype = tf.float16 if FLAGS.use_fp16 else tf.float32
var = _variable_on_cpu(
name,
shape,
tf.truncated_normal_initializer(stddev=stddev, dtype=dtype))

#计算weight decay项并加入到loss中。
if wd is not None:
weight_decay = tf.multiply(tf.nn.l2_loss(var), wd, name='weight_loss')
tf.add_to_collection('losses', weight_decay)
return var

调用 cifar10_input.distorted_inputs 来生成 distorted 的图像，用以扩充训练集。

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')
images, labels = cifar10_input.distorted_inputs(data_dir=data_dir,
batch_size=FLAGS.batch_size)
if FLAGS.use_fp16:
images = tf.cast(images, tf.float16)
labels = tf.cast(labels, tf.float16)
return images, labels

如果不想用 distorted 图像，调用 cifar10_input.inputs ，用来做测试而非训练用。

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')
images, labels = cifar10_input.inputs(eval_data=eval_data,
data_dir=data_dir,
batch_size=FLAGS.batch_size)
if FLAGS.use_fp16:
images = tf.cast(images, tf.float16)
labels = tf.cast(labels, tf.float16)
return images, labels

下面构建网络模型，第一层卷积层 tf.nn.conv2d

def inference(images):
"""Build the CIFAR-10 model.
Args:
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() wit
4000
h tf.Variable().
#
# conv1
with tf.variable_scope('conv1') as scope:
kernel = _variable_with_weight_decay('weights',
shape=[5, 5, 3, 64],
stddev=5e-2,
wd=0.0)
conv = tf.nn.conv2d(images, kernel, [1, 1, 1, 1], padding='SAME')
biases = _variable_on_cpu('biases', [64], tf.constant_initializer(0.0))
pre_activation = tf.nn.bias_add(conv, biases)
conv1 = tf.nn.relu(pre_activation, name=scope.name)
_activation_summary(conv1)

池化和归一化

# pool1
pool1 = tf.nn.max_pool(conv1, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1],
padding='SAME', name='pool1')
# norm1
norm1 = tf.nn.lrn(pool1, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75,
name='norm1')

下面撘层与上面相似，只不过在 local4 中卷积层的权重参数引入 weight decay。

# conv2
with tf.variable_scope('conv2') as scope:
kernel = _variable_with_weight_decay('weights',
shape=[5, 5, 64, 64],
stddev=5e-2,
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))
pre_activation = tf.nn.bias_add(conv, biases)
conv2 = tf.nn.relu(pre_activation, 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.
reshape = tf.reshape(pool2, [FLAGS.batch_size, -1])
dim = reshape.get_shape()[1].value
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)

输出 softmax 层，这里没有直接使用堆层函数是因为下面损失函数使用的是 tf.nn.sparse_softmax_cross_entropy_with_logits，该函数自行计算 softmax ，所以直接输出应作为 softmax 层输入的 tensor 。

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

构建计算损失的函数，用
cross entropy 来作为损失函数，同时加上需要 weight decay 的参数损失项。

def loss(logits, labels):
"""Add L2Loss to all the trainable variables.
Add summary 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.
"""
# Calculate the average cross entropy loss across the batch.
labels = tf.cast(labels, tf.int64)
cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(
labels=labels, logits=logits, name='cross_entropy_per_example')
cross_entropy_mean = tf.reduce_mean(cross_entropy, name='cross_entropy')
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).
return tf.add_n(tf.get_collection('losses'), name='total_loss')

下面的函数用来记录损失，而且对 tf.get_collection('losses') 和 total_loss 进行滑动均值处理，记录平滑后的值。该函数返回的是进行均值的 op 。

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.
loss_averages = tf.train.ExponentialMovingAverage(0.9, name='avg')
losses = tf.get_collection('losses')
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

下面的函数用来训练模型

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
decay_steps = int(num_batches_per_epoch * NUM_EPOCHS_PER_DECAY)

用
tf.train.exponential_decay 来对 learning rate 进行衰退。

# 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)

tf.control_dependencies 用来设置控制依赖项, 'with' 下面包含的 ops 只能在依赖项列表中的 ops 和 vars 都执行和计算完后才被执行。即 计算损失-计算梯度-更新参数。

# Compute gradients.ly
with tf.control_dependencies([loss_averages_op]):
opt = tf.train.GradientDescentOptimizer(lr)
grads = opt.compute_gradients(total_loss)

# Apply gradients.
apply_gradient_op = opt.apply_gradients(grads, global_step=global_step)

记录 vars/grads ，并且生成一个影子变量用来存储滑动均值后的变量值（variable_average.apply）。

# 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)
variables_averages_op = variable_averages.apply(tf.trainable_variables())

这里
tf.no_op 实际上是一个不做任何事情的 op，但是由于上面有依赖项，所以可以用它来控制 flow ，把它作为 train_op ，用以触发 梯度计算和参数更新来进行训练，但是他本身并不作任何事情。最终该训练函数返回一个用以训练(apply_grad, var_movave)  的op 。

with tf.control_dependencies([apply_gradient_op, variables_averages_op]):
train_op = tf.no_op(name='train')

return train_op