2016年直至现在,业内提出了SqueezeNet、ShuffleNet、NasNet、MnasNet以及MobileNet等轻量级网络模型。这些模型使移动终端、嵌入式设备运行神经网络模型成为可能。
V1(DW卷积)->V2(逆残差)->V3(SEattention)
1. 传统卷积
- 输入特征in_channel=卷积核的channel
- 输出特征图的out_channnel=卷积核的个数
计算量
与图像大小HW
、卷积核大小
有关
2. Mobilenet V1
2个亮点:
- DepthWise Separable卷积(深度可分离卷积)大大减少运算量:DW卷积+PW卷积
- 增加超参数alpha(控制卷积核个数)和beta(控制输入图像大小WH)
2.1 DepthWise Separable卷积
D
W
卷积
(
增大感受野,缩小特征图
W
H
)
+
P
W
卷积
(
缩放特征图通道数
C
)
DW卷积(增大感受野,缩小特征图WH)+PW卷积(缩放特征图通道数C)
DW卷积(增大感受野,缩小特征图WH)+PW卷积(缩放特征图通道数C)
Step1:逐通道独立卷积(DepthWise卷积):
一个卷积核负责一个通道
- 卷积核个数 = 输入特征图的in_channel = 输出特征图的out_channel
- 卷积核的channnel = 1
Step2:逐点卷积(Pointwise卷积)
卷积核1x1的普通卷积
,省参数,N个卷积核得到N个特征图
Step3:串联DW卷积和PW卷积
两步走的Separable卷积效果
与普通卷积
相似,但计算量大大减小
深度可分离卷积 VS 普通卷积:
2.2 整体结构
mobilenet v1一路卷积到底(单分支DW卷积块串行),但训练后DW卷积核很多参数为0,效果不理想。
3. Mobilenet V2
3.1 逆残差
- DW卷积:不改变通道数,但可缩小特征图WH
- 1x1卷积:不改变特征图WH,但可缩放通道数C
- 原始残差:先降维,后升维。两边厚,中间薄。
- 逆残差:先升维,后降维。两部薄,中间厚。
激活函数使用ReLU6:对ReLU设置max(6)
3.2 Relu干了坏事
Relu使 低维特征 损失信息
逆残差块输出
的是低位特征图(通道维度小),所以在输出时需要使用线性linear激活函数,避免信息损失
,而在逆残差块其他部分,仍然使用ReLU6激活函数。
3.3 整体结构
此处的bottleneck
就是逆残差块
,单分支一路卷积到底,bottleneck内部可能有shortcut。
4. Mobilenet V3
亮点:
- 更新bottleneck:引入
SE attention
,非线性激活函数NL将swish
改为h-swish
,用于替代relu6 - 使用
NAS
搜索参数(Neural Architecture Search) 重新设置耗时层
(对搜索出来的网络逐层分析优化)
4.1 更新bottleneck
①SE Attention:
Step1:求每个通道的权重值(1x1xc)(全局平均池化)
Step2:使用两个全连接层计算权重值的回归值sc(1x1xc),然后完成加权操作(每个通道权重回归值sc x 每个通道原始特征图uc)
②Switch激活函数:
基于Switch提出h-Switch激活函数,用于替代mobilenetV2中的relu6。
S
w
i
t
c
h
(
x
)
=
x
∗
s
i
g
m
o
i
d
(
x
)
Switch(x)=x*sigmoid(x)
Switch(x)=x∗sigmoid(x)
h
−
s
i
g
m
o
i
d
(
x
)
=
r
e
l
u
6
(
x
+
3
)
/
6
h-sigmoid(x)=relu6(x+3)/6
h−sigmoid(x)=relu6(x+3)/6
h
−
S
w
i
t
c
h
(
x
)
=
x
∗
h
−
s
i
g
m
o
i
d
(
x
)
=
x
∗
r
e
l
u
6
(
x
+
3
)
/
6
h-Switch(x)=x*h-sigmoid(x)=x*relu6(x+3)/6
h−Switch(x)=x∗h−sigmoid(x)=x∗relu6(x+3)/6
4.2 重新设置耗时层
因为实验发现第一个卷积操作
和最后的卷积操作
才是耗时性能瓶颈,所以减少了第一个卷积层的output channel从32到16
,且减少最后两层卷积层数量
。
4.3 整体结构
5. Mobilenet V3 代码
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.nn import init
class hswish(nn.Module):
def forward(self, x):
out = x * F.relu6(x + 3, inplace=True) / 6
return out
class hsigmoid(nn.Module):
def forward(self, x):
out = F.relu6(x + 3, inplace=True) / 6
return out
class SeModule(nn.Module):
def __init__(self, in_size, reduction=4):
super(SeModule, self).__init__()
self.se = nn.Sequential(
nn.AdaptiveAvgPool2d(1),
nn.Conv2d(in_size, in_size // reduction, kernel_size=1, stride=1, padding=0, bias=False),
nn.BatchNorm2d(in_size // reduction),
nn.ReLU(inplace=True),
nn.Conv2d(in_size // reduction, in_size, kernel_size=1, stride=1, padding=0, bias=False),
nn.BatchNorm2d(in_size),
hsigmoid()
)
def forward(self, x):
return x * self.se(x)
class Block(nn.Module):
'''expand + depthwise + pointwise'''
def __init__(self, kernel_size, in_size, expand_size, out_size, nolinear, semodule, stride):
super(Block, self).__init__()
self.stride = stride
self.se = semodule
self.conv1 = nn.Conv2d(in_size, expand_size, kernel_size=1, stride=1, padding=0, bias=False)
self.bn1 = nn.BatchNorm2d(expand_size)
self.nolinear1 = nolinear
self.conv2 = nn.Conv2d(expand_size, expand_size, kernel_size=kernel_size, stride=stride, padding=kernel_size//2, groups=expand_size, bias=False)
self.bn2 = nn.BatchNorm2d(expand_size)
self.nolinear2 = nolinear
self.conv3 = nn.Conv2d(expand_size, out_size, kernel_size=1, stride=1, padding=0, bias=False)
self.bn3 = nn.BatchNorm2d(out_size)
self.shortcut = nn.Sequential()
if stride == 1 and in_size != out_size:
self.shortcut = nn.Sequential(
nn.Conv2d(in_size, out_size, kernel_size=1, stride=1, padding=0, bias=False),
nn.BatchNorm2d(out_size),
)
def forward(self, x):
out = self.nolinear1(self.bn1(self.conv1(x)))
out = self.nolinear2(self.bn2(self.conv2(out)))
out = self.bn3(self.conv3(out))
if self.se != None:
out = self.se(out)
out = out + self.shortcut(x) if self.stride==1 else out
return out
class MobileNetV3_Large(nn.Module):
def __init__(self, num_classes=1000):
super(MobileNetV3_Large, self).__init__()
self.conv1 = nn.Conv2d(3, 16, kernel_size=3, stride=2, padding=1, bias=False)
self.bn1 = nn.BatchNorm2d(16)
self.hs1 = hswish()
self.bneck = nn.Sequential(
Block(3, 16, 16, 16, nn.ReLU(inplace=True), None, 1),
Block(3, 16, 64, 24, nn.ReLU(inplace=True), None, 2),
Block(3, 24, 72, 24, nn.ReLU(inplace=True), None, 1),
Block(5, 24, 72, 40, nn.ReLU(inplace=True), SeModule(40), 2),
Block(5, 40, 120, 40, nn.ReLU(inplace=True), SeModule(40), 1),
Block(5, 40, 120, 40, nn.ReLU(inplace=True), SeModule(40), 1),
Block(3, 40, 240, 80, hswish(), None, 2),
Block(3, 80, 200, 80, hswish(), None, 1),
Block(3, 80, 184, 80, hswish(), None, 1),
Block(3, 80, 184, 80, hswish(), None, 1),
Block(3, 80, 480, 112, hswish(), SeModule(112), 1),
Block(3, 112, 672, 112, hswish(), SeModule(112), 1),
Block(5, 112, 672, 160, hswish(), SeModule(160), 1),
Block(5, 160, 672, 160, hswish(), SeModule(160), 2),
Block(5, 160, 960, 160, hswish(), SeModule(160), 1),
)
self.conv2 = nn.Conv2d(160, 960, kernel_size=1, stride=1, padding=0, bias=False)
self.bn2 = nn.BatchNorm2d(960)
self.hs2 = hswish()
self.linear3 = nn.Linear(960, 1280)
self.bn3 = nn.BatchNorm1d(1280)
self.hs3 = hswish()
self.linear4 = nn.Linear(1280, num_classes)
self.init_params()
def init_params(self):
for m in self.modules():
if isinstance(m, nn.Conv2d):
init.kaiming_normal_(m.weight, mode='fan_out')
if m.bias is not None:
init.constant_(m.bias, 0)
elif isinstance(m, nn.BatchNorm2d):
init.constant_(m.weight, 1)
init.constant_(m.bias, 0)
elif isinstance(m, nn.Linear):
init.normal_(m.weight, std=0.001)
if m.bias is not None:
init.constant_(m.bias, 0)
def forward(self, x):
out = self.hs1(self.bn1(self.conv1(x)))
out = self.bneck(out)
out = self.hs2(self.bn2(self.conv2(out)))
out = F.avg_pool2d(out, 7)
out = out.view(out.size(0), -1)
out = self.hs3(self.bn3(self.linear3(out)))
out = self.linear4(out)
return out
class MobileNetV3_Small(nn.Module):
def __init__(self, num_classes=1000):
super(MobileNetV3_Small, self).__init__()
self.conv1 = nn.Conv2d(3, 16, kernel_size=3, stride=2, padding=1, bias=False)
self.bn1 = nn.BatchNorm2d(16)
self.hs1 = hswish()
self.bneck = nn.Sequential(
Block(3, 16, 16, 16, nn.ReLU(inplace=True), SeModule(16), 2),
Block(3, 16, 72, 24, nn.ReLU(inplace=True), None, 2),
Block(3, 24, 88, 24, nn.ReLU(inplace=True), None, 1),
Block(5, 24, 96, 40, hswish(), SeModule(40), 2),
Block(5, 40, 240, 40, hswish(), SeModule(40), 1),
Block(5, 40, 240, 40, hswish(), SeModule(40), 1),
Block(5, 40, 120, 48, hswish(), SeModule(48), 1),
Block(5, 48, 144, 48, hswish(), SeModule(48), 1),
Block(5, 48, 288, 96, hswish(), SeModule(96), 2),
Block(5, 96, 576, 96, hswish(), SeModule(96), 1),
Block(5, 96, 576, 96, hswish(), SeModule(96), 1),
)
self.conv2 = nn.Conv2d(96, 576, kernel_size=1, stride=1, padding=0, bias=False)
self.bn2 = nn.BatchNorm2d(576)
self.hs2 = hswish()
self.linear3 = nn.Linear(576, 1280)
self.bn3 = nn.BatchNorm1d(1280)
self.hs3 = hswish()
self.linear4 = nn.Linear(1280, num_classes)
self.init_params()
def init_params(self):
for m in self.modules():
if isinstance(m, nn.Conv2d):
init.kaiming_normal_(m.weight, mode='fan_out')
if m.bias is not None:
init.constant_(m.bias, 0)
elif isinstance(m, nn.BatchNorm2d):
init.constant_(m.weight, 1)
init.constant_(m.bias, 0)
elif isinstance(m, nn.Linear):
init.normal_(m.weight, std=0.001)
if m.bias is not None:
init.constant_(m.bias, 0)
def forward(self, x):
out = self.hs1(self.bn1(self.conv1(x)))
out = self.bneck(out)
out = self.hs2(self.bn2(self.conv2(out)))
out = F.avg_pool2d(out, 7)
out = out.view(out.size(0), -1)
out = self.hs3(self.bn3(self.linear3(out)))
out = self.linear4(out)
return out
def test():
net = MobileNetV3_Small()
x = torch.randn(2,3,224,224)
y = net(x)
print(y.size())
test()