VGGNet网络介绍
VGG系列结构图，
1，全部使用3*3的卷积核和2*2的池化核，通过不断加深网络结构来提升性能。
2，实验过程中使用了1*1的卷积，其意义在于线性变换，不改变通道数，也不起降维作用(1*1卷积的降维作用等在Inception中才得到展现)。实际上1*1的卷积也是有效的，但是器效果并不如3*3的好，大的卷积核可以学习大的空间特征
3，有五段卷积操作，每段feature_map数目相同，越靠后数目越多，分别为64-128-256-512-512
4、 Multi-scale训练
5，LRN(局部响应归一化层)作用不大
6，网络越深越好
VGG_19结构示意图，
VGGNet测试代码
注，网络版本为VGG_16
# Author : Hellcat
# Time : 2017/12/11 import math
import time
import tensorflow as tf
from datetime import datetime def conv_op(input_op, name, kh, kw, n_out, dh, dw, p):
'''
卷积层封装实现
核参数初始化使用了xavier方法
:param input_op:输入数据
:param name: 本层命名
:param kh: 卷积核高度
:param kw: 卷积核宽度
:param n_out: 输出 feature map 数量
:param dh: 高度方向步长
:param dw: 宽度方向步长
:param p: 变量收集器
:return: 输出层
'''
n_in = input_op.get_shape()[-1].value with tf.name_scope(name) as scope:
kernel = tf.get_variable(scope+'w',
shape=[kh,kw,n_in,n_out],dtype=tf.float32,
initializer=tf.contrib.layers.xavier_initializer_conv2d())
conv = tf.nn.conv2d(input_op,kernel,(1,dh,dw,1),padding='SAME')
bias_init_val = tf.constant(0.0, shape=[n_out],dtype=tf.float32)
biases = tf.Variable(bias_init_val,trainable=True,name='b')
z = tf.nn.bias_add(conv, biases)
activation = tf.nn.relu(z,name=scope)
p += [kernel, biases]
return activation def fc_op(input_op,name,n_out,p):
'''
全连接层封装实现
:param input_op:输入数据
:param name: 本层命名
:param n_out: 输出层节点数目
:param p: 变量收集器
:return: 输出层
'''
n_in = input_op.get_shape()[-1].value
with tf.name_scope(name) as scope:
kernel = tf.get_variable(scope+'w',
shape=[n_in,n_out],dtype=tf.float32,
initializer=tf.contrib.layers.xavier_initializer())
biases = tf.Variable(tf.constant(0.1,shape=[n_out],
dtype=tf.float32),name='b')
# relu_layer(x, weights, biases, name=None)
# relu(matmul(x, weights) + biases)
activation = tf.nn.relu_layer(input_op,kernel,biases,name=scope)
p += [kernel,biases]
return activation def mpool_op(input_op,name,kh,kw,dh,dw):
'''
池化层封装实现
:param input_op: 输入数据
:param name: 本层命名
:param kh: 核高度
:param kw: 核宽度
:param dh: 高度方向步长
:param dw: 宽度方向步长
:return: 输出层
'''
return tf.nn.max_pool(input_op,
ksize=[1,kh,kw,1],
strides=[1,dh,dw,1],
padding='SAME',
name=name) def inference_op(input_op, keep_prob):
'''
网络实现
:param input_op: 输入数据
:param keep_prob: dropout参数
:return: 输出结果
'''
p=[] conv1_1 = conv_op(input_op,name='conv1_1',kh=3,kw=3,n_out=64,dh=1,dw=1,p=p)
conv1_2 = conv_op(conv1_1,name='conv1_2',kh=3,kw=3,n_out=64,dh=1,dw=1,p=p)
pool1 = mpool_op(conv1_2,name='pool1',kh=2,kw=2,dh=2,dw=2) conv2_1 = conv_op(pool1,name='conv2_1',kh=3,kw=3,n_out=128,dh=1,dw=1,p=p)
conv2_2 = conv_op(conv2_1,name='conv2_2',kh=3,kw=3,n_out=128,dh=1,dw=1,p=p)
pool2 = mpool_op(conv2_2,name='pool2',kh=2,kw=2,dh=2,dw=2) conv3_1 = conv_op(pool2,name='conv3_1',kh=3,kw=3,n_out=256,dh=1,dw=1,p=p)
conv3_2 = conv_op(conv3_1,name='conv3_2',kh=3,kw=3,n_out=256,dh=1,dw=1,p=p)
conv3_3 = conv_op(conv3_2,name='conv3_3',kh=3,kw=3,n_out=256,dh=1,dw=1,p=p)
pool3 = mpool_op(conv3_3,name='pool3',kh=2,kw=2,dh=2,dw=2) conv4_1 = conv_op(pool3,name='conv4_1',kh=3,kw=3,n_out=512,dh=1,dw=1,p=p)
conv4_2 = conv_op(conv4_1,name='conv4_2',kh=3,kw=3,n_out=512,dh=1,dw=1,p=p)
conv4_3 = conv_op(conv4_2,name='conv4_3',kh=3,kw=3,n_out=512,dh=1,dw=1,p=p)
pool4 = mpool_op(conv4_3,name='pool4',kh=2,kw=2,dh=2,dw=2) conv5_1 = conv_op(pool4,name='conv5_1',kh=3,kw=3,n_out=512,dh=1,dw=1,p=p)
conv5_2 = conv_op(conv5_1,name='conv5_2',kh=3,kw=3,n_out=512,dh=1,dw=1,p=p)
conv5_3 = conv_op(conv5_2,name='conv5_3',kh=3,kw=3,n_out=512,dh=1,dw=1,p=p)
pool5 = mpool_op(conv5_3,name='pool5',kh=2,kw=2,dh=2,dw=2) shp = pool5.get_shape()
flattened_shape = shp[1].value * shp[2].value * shp[3].value
resh1 = tf.reshape(pool5,[-1,flattened_shape],name='resh1') fc6 = fc_op(resh1,name='fc6',n_out=4096,p=p)
fc6_drop = tf.nn.dropout(fc6,keep_prob=keep_prob,name='fc6_drop') fc7 = fc_op(fc6_drop,name='fc7',n_out=4096,p=p)
fc7_drop = tf.nn.dropout(fc7,keep_prob=keep_prob,name='fc7_drop') fc8 = fc_op(fc7_drop,name='fc8',n_out=1000,p=p)
softmax = tf.nn.softmax(fc8)
predictions = tf.argmax(softmax,axis=1) return predictions,softmax,fc8,p def time_tensorflow_run(session, target, feed, 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, feed_dict=feed)
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)) def run_benchmark():
with tf.Graph().as_default():
image_size = 224
images = tf.Variable(tf.random_normal([batch_size,image_size,image_size,3],
dtype=tf.float32,
stddev=1e-1))
keep_prob = tf.placeholder(tf.float32)
predictions,softmax,fc8,p = inference_op(images,keep_prob) init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init) time_tensorflow_run(sess,predictions,{keep_prob:1.0},'Forward')
objective = tf.nn.l2_loss(fc8) # 模拟目标函数
grad = tf.gradients(objective, p) # 梯度求解
time_tensorflow_run(sess, grad,{keep_prob:0.5},'Forward-backward') # 反向传播 batch_size = 32
num_batches = 100
run_benchmark()
输出如下面形式，我得机器实在是慢，进给出前几个输出示意，