网络结构分析
其中Inception模块组结构如下:
完整模型如下:
代码实现
1、导入模块
import tensorflow as tf
import tensorflow.contrib.slim as slim
from datetime import datetime
import math
import time
2、实现一个简单的函数trunc_normal,产生截断的正态分布
trunc_normal = lambda stddev: tf.truncated_normal_initializer(0.0, stddev)
3、定义函数inception_v3_arg_scope,用来生成网络中经常用到的函数的默认参数
def inception_v3_arg_scope(weight_decay=0.00004, stddev=0.1,
batch_norm_var_collection='moving_vars'):
"""生成网络中经常用到的函数的默认参数"""
batch_norm_params = {
'decay': 0.9997,
'epsilon': 0.001,
'updates_collections': tf.GraphKeys.UPDATE_OPS,
'variables_collections': {
'beta': None,
'gamma': None,
'moving_mean': [batch_norm_var_collection],
'moving_variance': [batch_norm_var_collection]
}
}
# slim.arg_scope可以给函数的参数自动赋予某些默认值,例如with slim.arg_scope([slim.conv2d,
# slim.fully_connected],weights_regularizer=slim.l2_regularizer(weight_decay))这句会对
# slim.conv2d, slim.fully_connected这两个函数的参数自动赋值,将参数weights_regularizer默认
# 设置为slim.l2_regularizer(weight_decay)。此后不需要每次都设置参数了只需要在修改的时候设置
with slim.arg_scope([slim.conv2d, slim.fully_connected],
weights_regularizer=slim.l2_regularizer(weight_decay)):
# 再对卷积层函数进行默认参数配置
with slim.arg_scope([slim.conv2d],
weights_initializer=tf.truncated_normal_initializer(stddev=stddev),
activation_fn=tf.nn.relu,
normalizer_fn=slim.batch_norm,
normalizer_params=batch_norm_params) as sc:
return sc
4、定义Inception_v3_base函数,用以生成Inception V3网络的卷积部分
def inception_v3_base(inputs, scope=None):
"""一个inception_v3基本模块"""
end_points = {} # 用来保存常用的关键节点
# qs1:这里的variable_scope的三个参数是什么意思
# 非Inception层的构建
with tf.variable_scope(scope, 'InceptionV3', [inputs]):
with slim.arg_scope([slim.conv2d, slim.max_pool2d, slim.avg_pool2d],
stride=1, padding='VALID'):
# 299 x 299 x 3
net = slim.conv2d(inputs, 32, [3, 3], stride=2, scope='Conv2d_1a_3x3')
# 149 x 149 x 32
net = slim.conv2d(net, 32, [3, 3], scope='Conv2d_2a_3x3')
# 147 x 147 x 32
net = slim.conv2d(net, 64, [3, 3], padding='SAME', scope='Conv2d_2b_3x3')
# 147 x 147 x 64
net = slim.max_pool2d(net, [3, 3], stride=2, scope='MaxPool_3a_3x3')
# 73 x 73 x 64
net = slim.conv2d(net, 80, [3, 3], scope='Conv2d_3b_1x1')
# 73 x 73 x 80.
net = slim.conv2d(net, 192, [3, 3], stride=2, scope='Conv2d_4a_3x3')
# 71 x 71 x 192.
net = slim.max_pool2d(net, [3, 3], scope='MaxPool_5a_3x3')
# 35 x 35 x 192.
with slim.arg_scope([slim.conv2d, slim.max_pool2d, slim.avg_pool2d],
stride=1, padding="SAME"):
# 第一个Inception模块组的第一个Inception模块构建
# 35x35x256
with tf.variable_scope('Mixed_5b'):
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], 3)
# 第一个模块组的第二个Inception模块——Mixed_5c构建
# 35x35x288
with tf.variable_scope('Mixed_5c'):
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='Conv_1_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')
# qs3:这里的3是什么意思
# an:是axis,轴的意思
net = tf.concat([branch_0, branch_1, branch_2, branch_3], 3)
# 第一个模块组的第三个模块——Mixed_5d和上一个模块一样
# 35x35x288
with tf.variable_scope('Mixed_5d'):
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='Conv_1_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], 3)
# 第二个模块组的第一个模块——Mixed_6a
# 17x17x768
with tf.variable_scope('Mixed_6a'):
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_1x1')
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)
# 第二个Inception模块组第二个Inception模块——Mixed_6b
# mixed4: 17 x 17 x 768.
with tf.variable_scope('Mixed_6b'):
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], 3)
# 第二个模块组的第二个Inception模块——Mixed_6c
# mixed_5: 17 x 17 x 768.
with tf.variable_scope('Mixed_6c'):
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], 3)
# mixed_6: 17 x 17 x 768.
with tf.variable_scope('Mixed_6d'):
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], 3)
# mixed_7: 17 x 17 x 768.
with tf.variable_scope('Mixed_6e'):
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], 3)
end_points['Mixed_6e'] = net
# 开始构筑第三个模块组
# mixed_8: 8 x 8 x 1280.
with tf.variable_scope('Mixed_7a'):
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)
# mixed_9: 8 x 8 x 2048.
with tf.variable_scope('Mixed_7b'):
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')], 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')], 3)
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], 3)
# mixed_10: 8 x 8 x 2048.
with tf.variable_scope('Mixed_7c'):
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_0c_3x1')], 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')], 3)
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], 3)
return net, end_points
5、生成Inception_v3网络的函数,其中添加了辅助分类层和最后最后的池化层,线性层和softmax层
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='InceptionV3'):
with tf.variable_scope(scope, 'InceptionV3',
[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)
# 辅助分类节点Auxiliaary Logits
with slim.arg_scope([slim.conv2d, slim.max_pool2d, slim.avg_pool2d],
stride=1, padding='SAME'):
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='SpatiaSqueeze')
end_points['AuxLogits'] = aux_logits
# 正常分类预测的逻辑
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='Predications')
return logits, end_points
6、定义测试用工具,可以测试每轮的训练时间
def time_tensorflow_run(session, target, info_string):
"""
评估Inception_V3每轮计算时间的函数
:param session: TensorFlow的Session
:param target: 需要预测的算子
:param info_string: 测试的名称
:return:
"""
num_steps_burn_in = 10 # 预热轮数,因为头几轮迭代有显存加载,cache命中等问题,所以不考虑
total_duration = 0.0 # 总时间
total_duration_squared = 0.0 # 总时间平方和
for i in range(num_batches + num_steps_burn_in):
start_time = time.time()
_ = session.run(target)
duration = time.time() - start_time
if i >= num_steps_burn_in: # 预热轮数之后再显示每轮消耗时间
if not i % 10:
print('%s: step %d, duration = %.3f' %
(datetime.now(), i - num_steps_burn_in, duration))
total_duration += duration # 累加用于计算后面每轮耗时的均值
total_duration_squared += duration * duration # 累加用于计算后面每轮函数的标准差
mn = total_duration / num_batches # 计算每轮平均耗时
vr = total_duration_squared / num_batches - mn * mn
sd = math.sqrt(vr) # 计算标准差
print('%s: %s across %d steps, %.3f +/- %.3f sec / batch' %
(datetime.now(), info_string, num_batches, mn, sd))
7、执行训练
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")
P.S.因为用的是笔记本CPU,所以只做了前向传播的部分,最终跑出来的速度并不好看。想要完全实现的可以自己添加损失函数和优化器,然后喂入数据就可以运行。
2018-08-13 12:08:11.923101: step 0, duration = 12.203
2018-08-13 12:10:13.557725: step 10, duration = 12.145
2018-08-13 12:12:15.377852: step 20, duration = 12.168
2018-08-13 12:14:17.043393: step 30, duration = 12.158
2018-08-13 12:16:18.598230: step 40, duration = 12.145
2018-08-13 12:18:20.088241: step 50, duration = 12.148
2018-08-13 12:20:21.695936: step 60, duration = 12.139
2018-08-13 12:22:23.915995: step 70, duration = 12.159
2018-08-13 12:24:25.987449: step 80, duration = 12.178
2018-08-13 12:26:28.586507: step 90, duration = 12.221
2018-08-13 12:28:18.476547: Forward across 100 steps, 12.188 +/- 0.067 sec / batch
8、InceptionV3总结
(1)Factorization into samll convolutions很有效,可以降低参数量,减轻过拟合,增加网络非线性的表达能力。
(2)卷积网络从输入到输出,应该让图片尺寸逐渐减小,输出通道数逐渐增加,即让空间结构简化,将空间信息转化为高阶抽象的特征信息。
(3)Inception Module用多个分支提取不同抽象程度的高阶特征的思路很有效,可以丰富网络的表达能力
9、问题总结
1、对slim.arg_scope的理解
slim.arg_scope可以给函数的参数自动赋予某些默认值,例如with slim.arg_scope([slim.conv2d,slim.fully_connected],weights_regularizer=slim.l2_regularizer(weight_decay))这句会对slim.conv2d, slim.fully_connected这两个函数的参数自动赋值,将参数weights_regularizer默认设置为slim.l2_regularizer(weight_decay)。此后不需要每次都设置参数了只需要在修改的时候设置。对于复杂网络的代码简化非常有效。
2、这里的variable_scope的三个参数是什么意思
with tf.variable_scope(scope, 'InceptionV3', [inputs]):
参考TensorFlow文档中
__init__(
name_or_scope,
default_name=None,
values=None,
.....
)
name_or_scope: string or VariableScope: the scope to open.
default_name: The default name to use if the name_or_scope argument is None, this name will be uniquified. If name_or_scope is provided it won't be used and therefore it is not required and can be None.
values: The list of Tensor arguments that are passed to the op function.
大意是第一个参数指定要打开的变量空间;
第二个参数是默认的该变量空间的名字,如果第一个参数不为None则第二个参数无效;
第三个参数是该变量空间中的函数需要用到的张量,必须显示得传到空间中。
3、为什么代码实现与书上结构对不上?
# 299 x 299 x 3
net = slim.conv2d(inputs, 32, [3, 3], stride=2, scope='Conv2d_1a_3x3')
# 149 x 149 x 32
net = slim.conv2d(net, 32, [3, 3], scope='Conv2d_2a_3x3')
# 147 x 147 x 32
net = slim.conv2d(net, 64, [3, 3], padding='SAME', scope='Conv2d_2b_3x3')
# 147 x 147 x 64
net = slim.max_pool2d(net, [3, 3], stride=2, scope='MaxPool_3a_3x3')
# 73 x 73 x 64
net = slim.conv2d(net, 80, [1, 1], scope='Conv2d_3b_1x1')
# 73 x 73 x 80.
net = slim.conv2d(net, 192, [3, 3], scope='Conv2d_4a_3x3')
# 71 x 71 x 192.
net = slim.max_pool2d(net, [3, 3], stride=2, scope='MaxPool_5a_3x3')
# 35 x 35 x 192.
书中代码错误,改为:
# 299 x 299 x 3
net = slim.conv2d(inputs, 32, [3, 3], stride=2, scope='Conv2d_1a_3x3')
# 149 x 149 x 32
net = slim.conv2d(net, 32, [3, 3], scope='Conv2d_2a_3x3')
# 147 x 147 x 32
net = slim.conv2d(net, 64, [3, 3], padding='SAME', scope='Conv2d_2b_3x3')
# 147 x 147 x 64
net = slim.max_pool2d(net, [3, 3], stride=2, scope='MaxPool_3a_3x3')
# 73 x 73 x 64
net = slim.conv2d(net, 80, [3, 3], scope='Conv2d_3b_1x1')
# 73 x 73 x 80.
net = slim.conv2d(net, 192, [3, 3], stride=2, scope='Conv2d_4a_3x3')
# 71 x 71 x 192.
net = slim.max_pool2d(net, [3, 3], scope='MaxPool_5a_3x3')
# 35 x 35 x 192.
4、这里的3是什么意思,怎么定义的
net = tf.concat([branch_0, branch_1, branch_2, branch_3], 3)
3是axis,轴的意思。根据Inception模块的分支数来决定,有时候为了保证维度的一致性会产生空轴,最后再去除。