『TensorFlow』读书笔记_Inception_V3_下
极为庞大的网络结构，不过下一节的ResNet也不小
线性的组成，结构大体如下：
常规卷积部分->Inception模块组1->Inception模块组2->Inception模块组3->池化->1*1卷积(实现个线性变换)->分类器
|_>辅助分类器
代码如下，
# Author : Hellcat
# Time : 2017/12/12
# refer : https://github.com/tensorflow/models/
# blob/master/research/inception/inception/slim/inception_model.py
import time
import math
import tensorflow as tf
from datetime import datetime
slim = tf.contrib.slim
# 截断误差初始化生成器
trunc_normal = lambda stddev:tf.truncated_normal_initializer(0.0,stddev)
def inception_v3_arg_scope(weight_decay=0.00004,
stddv=0.1,
batch_norm_var_collection='moving_vars'):
'''
网络常用函数默认参数生成
:param weight_decay: L2正则化decay
:param stddv: 标准差
:param batch_norm_var_collection:
:return:
'''
batch_norm_params = {
'decay':0.9997, # 衰减系数
'epsilon':0.001,
'updates_collections':{
'bate':None,
'gamma':None,
'moving_mean':[batch_norm_var_collection], # 批次均值
'moving_variance':[batch_norm_var_collection] # 批次方差
}
}
# 外层环境
with slim.arg_scope([slim.conv2d,slim.fully_connected],
# 权重正则化函数
weights_regularizer=slim.l2_regularizer(weight_decay)):
# 内层环境
with slim.arg_scope([slim.conv2d],
# 权重初始化函数
weights_initializer=tf.truncated_normal_initializer(stddev=stddv),
# 激活函数，默认为nn.relu
activation_fn=tf.nn.relu,
# 正则化函数，默认为None
normalizer_fn=slim.batch_norm,
# 正则化函数参数，字典形式
normalizer_params=batch_norm_params) as sc:
return sc
def inception_v3_base(inputs,scope=None):
# 保存关键节点
end_points = {}
# 重载作用域的名称，创建新的作用域名称(前面是None时使用)，输入tensor
with tf.variable_scope(scope,'Inception_v3',[inputs]):
with slim.arg_scope([slim.conv2d,slim.max_pool2d,slim.avg_pool2d],
stride=1,padding='VALID'):
# 299*299*3
net = slim.conv2d(inputs,32,[3,3],stride=2,scope='Conv2d_1a_3x3') # 149*149*32
net = slim.conv2d(net,32,[3,3],scope='Conv2d_2a_3x3') # 147*147*32
net = slim.conv2d(net,64,[3,3],padding='SAME',scope='Conv2d_2b_3x3') # 147*147*64
net = slim.max_pool2d(net,[3,3],stride=2,scope='MaxPool_3a_3x3') # 73*73*64
net = slim.conv2d(net,80,[1,1],scope='Conv2d_3b_1x1') # 73*73*80
net = slim.conv2d(net,192,[1,1],scope='Conv2d_4a_3x3') # 71*71*192
net = slim.max_pool2d(net,[3,3],stride=2,scope='MaxPool_5a_3x3') # 35*35*192
with slim.arg_scope([slim.conv2d,slim.max_pool2d,slim.avg_pool2d],
stride=1,padding='SAME'):
'''Inception 第一模组块'''
# Inception_Module_1
with tf.variable_scope('Mixed_5b'): # 35*35*256
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,64,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,48,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,64,[5,5],scope='Conv2d_0b_5x5')
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,64,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,96,[3,3],scope='Conv2d_0b_3x3')
branch_2 = slim.conv2d(branch_2,96,[3,3],scope='Conv2d_0c_3x3')
with tf.variable_scope('Branch_3'):
branch_3 = slim.avg_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,32,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],axis=3)
# Inception_Module_2
with tf.variable_scope('Mixed_5c'): # 35*35*288
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,64,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,48,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,64,[5,5],scope='Conv2d_0b_5x5')
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,64,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,96,[3,3],scope='Conv2d_0b_3x3')
branch_2 = slim.conv2d(branch_2,96,[3,3],scope='Conv2d_0c_3x3')
with tf.variable_scope('Branch_3'):
branch_3 = slim.avg_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,64,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],axis=3)
# Inception_Module_3
with tf.variable_scope('Mixed_5d'): # 35*35*288
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,64,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,48,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,64,[5,5],scope='Conv2d_0b_5x5')
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,64,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,96,[3,3],scope='Conv2d_0b_3x3')
branch_2 = slim.conv2d(branch_2,96,[3,3],scope='Conv2d_0c_3x3')
with tf.variable_scope('Branch_3'):
branch_3 = slim.avg_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,64,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],axis=3)
'''Inception 第二模组块'''
# Inception_Module_1
with tf.variable_scope('Mixed_6a'): # 17*17*768
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,384,[3,3],stride=2,
padding='VALID',scope='Conv2d_1a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,64,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,96,[3,3],scope='Conv2d_0b_3x3')
branch_1 = slim.conv2d(branch_1,96,[3,3],stride=2,
padding='VALID',scope='Conv2d_1a_3x3')
with tf.variable_scope('Branch_2'):
branch_2 = slim.max_pool2d(net,[3,3],stride=2,padding='VALID',
scope='Max_Pool_1a_3x3')
net = tf.concat([branch_0,branch_1,branch_2],axis=3)
# Inception_Module_2
with tf.variable_scope('Mixed_6b'): # 17*17*768
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,128,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,128,[1,7],scope='Conv2d_0b_1x7')
branch_1 = slim.conv2d(branch_1,192,[7,1],scope='Conv2d_0c_7x1')
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,128,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,128,[7,1],scope='Conv2d_0b_7x1')
branch_2 = slim.conv2d(branch_2,128,[1,7],scope='Conv2d_0c_1x7')
branch_2 = slim.conv2d(branch_2,128,[7,1],scope='Conv2d_0d_7x1')
branch_2 = slim.conv2d(branch_2,192,[1,7],scope='Conv2d_0e_1x7')
with tf.variable_scope('Branch_3'):
branch_3 = slim.avg_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,192,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],axis=3)
# Inception_Module_3
with tf.variable_scope('Mixed_6c'): # 17*17*768
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,160,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,160,[1,7],scope='Conv2d_0b_1x7')
branch_1 = slim.conv2d(branch_1,192,[7,1],scope='Conv2d_0c_7x1')
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,160,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,160,[7,1],scope='Conv2d_0b_7x1')
branch_2 = slim.conv2d(branch_2,160,[1,7],scope='Conv2d_0c_1x7')
branch_2 = slim.conv2d(branch_2,160,[7,1],scope='Conv2d_0d_7x1')
branch_2 = slim.conv2d(branch_2,192,[1,7],scope='Conv2d_0e_1x7')
with tf.variable_scope('Branch_3'):
branch_3 = slim.avg_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,192,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],axis=3)
# Inception_Module_4
with tf.variable_scope('Mixed_6d'): # 17*17*768
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,160,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,160,[1,7],scope='Conv2d_0b_1x7')
branch_1 = slim.conv2d(branch_1,192,[7,1],scope='Conv2d_0c_7x1')
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,160,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,160,[7,1],scope='Conv2d_0b_7x1')
branch_2 = slim.conv2d(branch_2,160,[1,7],scope='Conv2d_0c_1x7')
branch_2 = slim.conv2d(branch_2,160,[7,1],scope='Conv2d_0d_7x1')
branch_2 = slim.conv2d(branch_2,192,[1,7],scope='Conv2d_0e_1x7')
with tf.variable_scope('Branch_3'):
branch_3 = slim.avg_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,192,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],axis=3)
# Inception_Module_5
with tf.variable_scope('Mixed_6e'): # 17*17*768
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,192,[1,7],scope='Conv2d_0b_1x7')
branch_1 = slim.conv2d(branch_1,192,[7,1],scope='Conv2d_0c_7x1')
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,192,[7,1],scope='Conv2d_0b_7x1')
branch_2 = slim.conv2d(branch_2,192,[1,7],scope='Conv2d_0c_1x7')
branch_2 = slim.conv2d(branch_2,192,[7,1],scope='Conv2d_0d_7x1')
branch_2 = slim.conv2d(branch_2,192,[1,7],scope='Conv2d_0e_1x7')
with tf.variable_scope('Branch_3'):
branch_3 = slim.avg_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,192,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],axis=3)
end_points['Mixed_6e'] = net
'''Inception 第三模组块'''
# Inception_Module_1
with tf.variable_scope('Mixed_7a'): # 8*8*1280
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
branch_0 = slim.conv2d(branch_0,320,[3,3],stride=2,
padding='VALID',scope='Conv2d_1a_3x3')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,192,[1,1],scope='Conv2d_0a_1x1')
branch_1 = slim.conv2d(branch_1,192,[1,7],scope='Conv2d_0b_1x7')
branch_1 = slim.conv2d(branch_1,192,[7,1],scope='Conv2d_0c_7x1')
branch_1 = slim.conv2d(branch_1,192,[3,3],stride=2,padding='VALID',scope='Conv2d_1a_3x3')
with tf.variable_scope('Branch_2'):
branch_2 = slim.max_pool2d(net,[3,3],stride=2,padding='VALID',
scope='MaxPool_1a_3x3')
net = tf.concat([branch_0,branch_1,branch_2],3)
# Inception_Module_2
with tf.variable_scope('Mixed_7b'): # 8*8*2048
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,320,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,384,[1,1],scope='Conv2d_0a_1x1')
branch_1 = tf.concat([
slim.conv2d(branch_1,384,[1,3],scope='Conv2d_0b_1x3'),
slim.conv2d(branch_1,384,[3,1],scope='Conv2d_0b_3x1')],axis=3)
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,448,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,384,[3,3],scope='Conv2d_0b_3x3')
branch_2 = tf.concat([
slim.conv2d(branch_2,384,[1,3],scope='Conv2d_0c_1x3'),
slim.conv2d(branch_2,384,[3,1],scope='Conv2d_0d_3x1')],axis=3)
with tf.variable_scope('Branch_3'):
branch_3 = slim.max_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,192,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],3)
# Inception_Module_3
with tf.variable_scope('Mixed_7c'): # 8*8*2048
with tf.variable_scope('Branch_0'):
branch_0 = slim.conv2d(net,320,[1,1],scope='Conv2d_0a_1x1')
with tf.variable_scope('Branch_1'):
branch_1 = slim.conv2d(net,384,[1,1],scope='Conv2d_0a_1x1')
branch_1 = tf.concat([
slim.conv2d(branch_1,384,[1,3],scope='Conv2d_0b_1x3'),
slim.conv2d(branch_1,384,[3,1],scope='Conv2d_0b_3x1')],axis=3)
with tf.variable_scope('Branch_2'):
branch_2 = slim.conv2d(net,448,[1,1],scope='Conv2d_0a_1x1')
branch_2 = slim.conv2d(branch_2,384,[3,3],scope='Conv2d_0b_3x3')
branch_2 = tf.concat([
slim.conv2d(branch_2,384,[1,3],scope='Conv2d_0c_1x3'),
slim.conv2d(branch_2,384,[3,1],scope='Conv2d_0d_3x1')],axis=3)
with tf.variable_scope('Branch_3'):
branch_3 = slim.max_pool2d(net,[3,3],scope='AvgPool_0a_3x3')
branch_3 = slim.conv2d(branch_3,192,[1,1],scope='Conv2d_0b_1x1')
net = tf.concat([branch_0,branch_1,branch_2,branch_3],3)
return net,end_points
def inception_v3(inputs,
num_classes=1000,
is_training=True,
dropout_keep_prob=0.8,
prediction_fn=slim.softmax,
spatial_squeeze=True,
reuse=None,
scope='Inception_v3'):
with tf.variable_scope(scope,'Inception_v3',[inputs,num_classes],reuse=reuse) as scope:
with slim.arg_scope([slim.batch_norm,slim.dropout],
is_training=is_training):
net,end_points = inception_v3_base(inputs,scope=scope)
with slim.arg_scope([slim.conv2d,slim.max_pool2d,slim.avg_pool2d],
stride=1,padding='SAME'):
# 17*17*768
aux_logits = end_points['Mixed_6e']
with tf.variable_scope('AuxLogits'):
aux_logits = slim.avg_pool2d(aux_logits,[5,5],stride=3,padding='VALID',scope='AvgPool_1a_5x5')
aux_logits = slim.conv2d(aux_logits,128,[1,1],scope='Conv2d_1b_1x1')
aux_logits = slim.conv2d(aux_logits,768,[5,5],
weights_initializer=trunc_normal(0.01),
padding='VALID',
scope='Conv2d_2a_5x5')
aux_logits = slim.conv2d(aux_logits,num_classes,[1,1],activation_fn=None,
normalizer_fn=None,weights_initializer=trunc_normal(0.001),
scope='Conv2d_2b_1x1')
if spatial_squeeze:
aux_logits = tf.squeeze(aux_logits,[1,2],
name='SpatialSqueeze')
end_points['AuxLogits'] = aux_logits
with tf.variable_scope('Logits'):
net = slim.avg_pool2d(net,[8,8],padding='VALID',
scope='AvgPool_1a_8x8')
net = slim.dropout(net,keep_prob=dropout_keep_prob,scope='Dropout_1b')
end_points['PreLogits'] = net
logits = slim.conv2d(net,num_classes,[1,1],activation_fn=None,
normalizer_fn=None,scope='Conv2d_1c_1x1')
if spatial_squeeze:
logits = tf.squeeze(logits,[1,2],name='SpatialSqueeze')
end_points['Logits'] = logits
end_points['Predictions'] = prediction_fn(logits,scope='Predictions')
return logits, end_points
def time_tensorflow_run(session, target, info_string):
'''
网路运行时间测试函数
:param session: 会话对象
:param target: 运行目标节点
:param info_string:提示字符
:return: None
'''
num_steps_burn_in = 10 # 预热轮数
total_duration = 0.0 # 总时间
total_duration_squared = 0.0 # 总时间平方和
for i in range(num_steps_burn_in + num_batches):
start_time = time.time()
_ = session.run(target)
duration = time.time() - start_time # 本轮时间
if i >= num_steps_burn_in:
if not i % 10:
print('%s: step %d, duration = %.3f' %
(datetime.now(),i-num_steps_burn_in,duration))
total_duration += duration
total_duration_squared += duration**2
mn = total_duration/num_batches # 平均耗时
vr = total_duration_squared/num_batches - mn**2
sd = math.sqrt(vr)
print('%s:%s across %d steps, %.3f +/- %.3f sec / batch' %
(datetime.now(), info_string, num_batches, mn, sd))
if __name__ == '__main__':
batch_size=32
height,width = 299,299
inputs = tf.random_uniform((batch_size,height,width,3))
with slim.arg_scope(inception_v3_arg_scope()):
logits,end_points = inception_v3(inputs,is_training=False)
init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init)
num_batches = 100
time_tensorflow_run(sess,logits,'Forward')
运行起来时耗过长，就不贴了。
