Ghost Net

1. Introduction

上图是由ResNet-50中的第一个残差块生成的某些中间特征图的可视化。从图中我们可以看出，这里面有很多特征图是具有高度相似性的（在图中分别用不同的颜色示意），换句话说，就是存在许多的冗余特征图。所以从另一个角度想，我们是不是可以利用一系列的线性变化，以很小的代价生成许多能从原始特征发掘所需信息的“幻影”特征图呢？这个便是整篇文章的核心思想。

2. Approach

Ghost module：

先通过conv生成一些特征图
然后对生成的特征图进行cheap操作（Dwise conv）生成冗余特征图
最后将conv生成的特征图与cheap操作生成的特征图进行concat操作

3. GhostNet

3.1 Ghost Bottlenecks

第一层Ghost module充当扩展层用于增加通道的数目

第二层Ghost module减少通道的数目来满足shortcut path

3.2 网络结构

4. Pytorch实现

import torch
import torch.nn as nnclass DWConv3x3BNReLU(nn.Sequential):def __init__(self, in_channel, out_channel, stride, groups):super(DWConv3x3BNReLU, self).__init__(nn.Conv2d(in_channels=in_channel, out_channels=out_channel, kernel_size=3, stride=stride, padding=1, groups=groups, bias=False),nn.BatchNorm2d(out_channel),nn.ReLU6(inplace=True),)class SqueezeAndExcite(nn.Module):def __init__(self, in_channel, out_channel, divide=4):super(SqueezeAndExcite, self).__init__()mid_channel = in_channel // divideself.pool = nn.AdaptiveAvgPool2d((1, 1))self.SEblock = nn.Sequential(nn.Linear(in_features=in_channel, out_features=mid_channel),nn.ReLU6(inplace=True),nn.Linear(in_features=mid_channel, out_features=out_channel),nn.ReLU6(inplace=True),)def forward(self, x):b, c, h, w = x.size()out = self.pool(x)out = torch.flatten(out, start_dim=1)out = self.SEblock(out)out = out.view(b, c, 1, 1)return out * xclass GhostModule(nn.Module):def __init__(self, in_channel, out_channel, s=2, kernel_size=1, stride=1, use_relu=True):super(GhostModule, self).__init__()intrinsic_channel = out_channel // sghost_channel = intrinsic_channel * (s - 1)self.primary_conv = nn.Sequential(nn.Conv2d(in_channels=in_channel, out_channels=intrinsic_channel, kernel_size=kernel_size, stride=stride, padding=(kernel_size-1)//2, bias=False),nn.BatchNorm2d(intrinsic_channel),nn.ReLU(inplace=True) if use_relu else nn.Sequential())self.cheap_op = DWConv3x3BNReLU(in_channel=intrinsic_channel, out_channel=ghost_channel, stride=stride, groups=intrinsic_channel)def forward(self, x):x1 = self.primary_conv(x)x2 = self.cheap_op(x1)out = torch.cat([x1, x2], dim=1)return outclass GhostBottleneck(nn.Module):def __init__(self, in_channel, mid_channel, out_channel, kernel_size, stride, use_se):super(GhostBottleneck, self).__init__()self.stride = strideself.bottleneck = nn.Sequential(GhostModule(in_channel=in_channel, out_channel=mid_channel, use_relu=True),DWConv3x3BNReLU(in_channel=mid_channel, out_channel=mid_channel, stride=stride, groups=mid_channel) if self.stride > 1 else nn.Sequential(),SqueezeAndExcite(in_channel=mid_channel, out_channel=mid_channel) if use_se else nn.Sequential(),GhostModule(in_channel=mid_channel, out_channel=out_channel, use_relu=False))if self.stride > 1:self.shortcut = DWConv3x3BNReLU(in_channel=in_channel, out_channel=out_channel, stride=stride, groups=1)else:self.shortcut = nn.Conv2d(in_channels=in_channel, out_channels=out_channel, kernel_size=1, stride=1)def forward(self, x):out = self.bottleneck(x)residual = self.shortcut(x)out += residualreturn outclass GhostNet(nn.Module):def __init__(self, num_classes=1000):super(GhostNet, self).__init__()self.first_conv = nn.Sequential(nn.Conv2d(in_channels=3, out_channels=16, kernel_size=3, stride=2, padding=1, bias=False),nn.BatchNorm2d(16),nn.ReLU6(inplace=True),)ghost_model_setting = [# in, mid, out, kernel, stride, use_se[16, 16, 16, 3, 1, False],[16, 48, 24, 3, 2, False],[24, 72, 24, 3, 1, False],[24, 72, 40, 5, 2, True],[40, 120, 40, 5, 1, True],[40, 240, 80, 3, 2, False],[80, 200, 80, 3, 1, False],[80, 184, 80, 3, 1, False],[80, 184, 80, 3, 1, False],[80, 480, 112, 3, 1, True],[112, 672, 112, 3, 1, True],[112, 672, 160, 5, 2, True],[160, 960, 160, 5, 1, False],[160, 960, 160, 5, 1, True],[160, 960, 160, 5, 1, False],[160, 960, 160, 5, 1, True],]layers = []for in_channel, mid_channel, out_channel, kernel_size, stride, use_se in ghost_model_setting:layers.append(GhostBottleneck(in_channel=in_channel, mid_channel=mid_channel, out_channel=out_channel, kernel_size=kernel_size, stride=stride, use_se=use_se))self.features = nn.Sequential(*layers)self.last_stage = nn.Sequential(nn.Conv2d(in_channels=160, out_channels=960, kernel_size=1, stride=1),nn.BatchNorm2d(960),nn.AdaptiveAvgPool2d((1, 1)),nn.Conv2d(in_channels=960, out_channels=1280, kernel_size=1, stride=1),nn.ReLU6(inplace=True))self.classifier = nn.Linear(in_features=1280, out_features=num_classes)def forward(self, x):x = self.first_conv(x)x = self.features(x)x = self.last_stage(x)x = torch.flatten(x, start_dim=1)x = self.classifier(x)return xmodel = GhostNet()
input = torch.randn(1, 3, 224, 224)
out = model(input)
print(out.size())

上一篇：ShuffleNet系列

完整代码

GhostNet详解及代码实现相关推荐

调包侠福音！机器学习经典算法开源教程（附参数详解及代码实现）
Datawhale 作者:赵楠.杨开漠.谢文昕.张雨寄语:本文针对5大机器学习经典算法,梳理了其模型.策略和求解等方面的内容,同时给出了其对应sklearn的参数详解和代码实现,帮助学习者入门和巩固 ...
粒子群(pso)算法详解matlab代码,粒子群(pso)算法详解matlab代码
粒子群(pso)算法详解matlab代码 (1)---- 一.粒子群算法的历史粒子群算法源于复杂适应系统(Complex Adaptive System,CAS).CAS理论于1994年正式提出,C ...
图像质量损失函数SSIM Loss的原理详解和代码具体实现
本文转自微信公众号SIGAI 文章PDF见: http://www.tensorinfinity.com/paper_164.html http://www.360doc.com/content/19 ...
python 自动化-Python API 自动化实战详解(纯代码)
主要讲如何在公司利用Python 搞API自动化. 1.分层设计思路 dataPool :数据池层,里面有我们需要的各种数据,包括一些公共数据等 config :基础配置 tools : 工具层 co ...
数学建模——智能优化之遗传算法详解Python代码
数学建模--智能优化之遗传算法详解Python代码 import numpy as np import matplotlib.pyplot as plt from matplotlib import ...
数学建模——主成分分析算法详解Python代码
数学建模--主成分分析算法详解Python代码 import matplotlib.pyplot as plt #加载matplotlib用于数据的可视化 from sklearn.decomposi ...
数学建模——智能优化之模拟退火模型详解Python代码
数学建模--智能优化之模拟退火模型详解Python代码 #本功能实现最小值的求解#from matplotlib import pyplot as plt import numpy as np imp ...
数学建模——智能优化之粒子群模型详解Python代码
数学建模--智能优化之粒子群模型详解Python代码 import numpy as np import matplotlib.pyplot as plt from mpl_toolkits.mplo ...
数学建模——支持向量机模型详解Python代码
数学建模--支持向量机模型详解Python代码 from numpy import * import random import matplotlib.pyplot as plt import num ...

GhostNet详解及代码实现