提示：文章写完后，目录可以自动生成，如何生成可参考右边的帮助文档
文章目录
前言
本文主要介绍一种在YOLOv5-7.0中添加EMA注意力机制的方法。EMA注意力机制原论文地址，有关EMA注意力机制的解读可参考文章。
新建EMA.py文件
在yolov5的models文件中新建一个名为EMA.py文件，将下述代码复制到EMA.py文件中并保存。
import torchfrom torch import nnclass EMA(nn.Module): def __init__(self, channels, factor=8): super(EMA, self).__init__() self.groups = factor assert channels // self.groups > 0 self.softmax = nn.Softmax(-1) self.agp = nn.AdaptiveAvgPool2d((1, 1)) self.pool_h = nn.AdaptiveAvgPool2d((None, 1)) self.pool_w = nn.AdaptiveAvgPool2d((1, None)) self.gn = nn.GroupNorm(channels // self.groups, channels // self.groups) self.conv1x1 = nn.Conv2d(channels // self.groups, channels // self.groups, kernel_size=1, stride=1, padding=0) self.conv3x3 = nn.Conv2d(channels // self.groups, channels // self.groups, kernel_size=3, stride=1, padding=1) def forward(self, x): b, c, h, w = x.size() group_x = x.reshape(b * self.groups, -1, h, w) # b*g,c//g,h,w x_h = self.pool_h(group_x) x_w = self.pool_w(group_x).permute(0, 1, 3, 2) hw = self.conv1x1(torch.cat([x_h, x_w], dim=2)) x_h, x_w = torch.split(hw, [h, w], dim=2) x1 = self.gn(group_x * x_h.sigmoid() * x_w.permute(0, 1, 3, 2).sigmoid()) x2 = self.conv3x3(group_x) x11 = self.softmax(self.agp(x1).reshape(b * self.groups, -1, 1).permute(0, 2, 1)) x12 = x2.reshape(b * self.groups, c // self.groups, -1) # b*g, c//g, hw x21 = self.softmax(self.agp(x2).reshape(b * self.groups, -1, 1).permute(0, 2, 1)) x22 = x1.reshape(b * self.groups, c // self.groups, -1) # b*g, c//g, hw weights = (torch.matmul(x11, x12) + torch.matmul(x21, x22)).reshape(b * self.groups, 1, h, w) return (group_x * weights.sigmoid()).reshape(b, c, h, w)修改yolo.py文件
1.导入EMA.py
在yolo.py文件开头导入EMA.py，代码如下：
from models.EMA import EMA代码放在yolo.py位置如下图所示：
2.修改parse_model
这里主要是添加通道参数，再添加一个elif，把EMA添加进去，代码如下：
elif m is EMA: args = [ch[f], *args]添加上述代码的位置可参考下图：
修改yaml文件（yolov5s为例）
# YOLOv5 🚀 by Ultralytics, GPL-3.0 license# Parametersnc: 80 # number of classesdepth_multiple: 0.33 # model depth multiplewidth_multiple: 0.50 # layer channel multipleanchors: - [10,13, 16,30, 33,23] # P3/8 - [30,61, 62,45, 59,119] # P4/16 - [116,90, 156,198, 373,326] # P5/32# YOLOv5 v6.0 backbonebackbone: # [from, number, module, args] [[-1, 1, Conv, [64, 6, 2, 2]], # 0-P1/2 [-1, 1, Conv, [128, 3, 2]], # 1-P2/4 [-1, 3, C3, [128]], [-1, 1, Conv, [256, 3, 2]], # 3-P3/8 [-1, 6, C3, [256]], [-1, 1, Conv, [512, 3, 2]], # 5-P4/16 [-1, 9, C3, [512]], [-1, 1, Conv, [1024, 3, 2]], # 7-P5/32 [-1, 3, C3, [1024]], [-1, 1, EMA, [8]], [-1, 1, SPPF, [1024, 5]], # 9 ]# YOLOv5 v6.0 headhead: [[-1, 1, Conv, [512, 1, 1]], [-1, 1, nn.Upsample, [None, 2, 'nearest']], [[-1, 6], 1, Concat, [1]], # cat backbone P4 [-1, 3, C3, [512, False]], # 13 [-1, 1, Conv, [256, 1, 1]], [-1, 1, nn.Upsample, [None, 2, 'nearest']], [[-1, 4], 1, Concat, [1]], # cat backbone P3 [-1, 3, C3, [256, False]], # 17 (P3/8-small) [-1, 1, Conv, [256, 3, 2]], [[-1, 15], 1, Concat, [1]], # cat head P4 [-1, 3, C3, [512, False]], # 20 (P4/16-medium) [-1, 1, Conv, [512, 3, 2]], [[-1, 11], 1, Concat, [1]], # cat head P5 [-1, 3, C3, [1024, False]], # 23 (P5/32-large) [[18, 21, 24], 1, Detect, [nc, anchors]], # Detect(P3, P4, P5) ]上述代码将EMA注意力机制模块加在backbone层中最后C3模块后面，SPPF模块前面，仅供参考，具体添加位置要根据个人数据集的不同合理的添加。
[-1, 1, EMA, [8]], #-1代表连接上一层通道数，1是个数，8是EMA所需的参数（factor=8）说明：因为在yolo.py文件parse_model函数中修改了通道参数，因此在yaml文件中无需添加通道参数，只需添加EMA函数所需的其他参数。在backbone中添加一层注意力机制模块，因此后续的层数都要加一，在head层中做如下改动。
[[-1, 15], 1, Concat, [1]], #未改动前的第14层，在经过上述改动后改为15[[-1, 11], 1, Concat, [1]], #未改动前的第10层，在记过上述改动后改为11[[18, 21, 24], 1, Detect, [nc, anchors]], #17，20，23层改为18，21，24运行train.py文件可以在输出终端窗口看到上图网络结构，可以看到在第9层已经成功添加EMA注意力机制模块。
from n params module arguments 0 -1 1 3520 models.common.Conv [3, 32, 6, 2, 2] 1 -1 1 18560 models.common.Conv [32, 64, 3, 2] 2 -1 1 18816 models.common.C3 [64, 64, 1] 3 -1 1 73984 models.common.Conv [64, 128, 3, 2] 4 -1 2 115712 models.common.C3 [128, 128, 2] 5 -1 1 295424 models.common.Conv [128, 256, 3, 2] 6 -1 3 625152 models.common.C3 [256, 256, 3] 7 -1 1 1180672 models.common.Conv [256, 512, 3, 2] 8 -1 1 1182720 models.common.C3 [512, 512, 1] 9 -1 1 41216 models.EMA.EMA [512, 8] 10 -1 1 656896 models.common.SPPF [512, 512, 5] 11 -1 1 131584 models.common.Conv [512, 256, 1, 1] 12 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 13 [-1, 6] 1 0 models.common.Concat [1]14 -1 1 361984 models.common.C3 [512, 256, 1, False] 15 -1 1 33024 models.common.Conv [256, 128, 1, 1] 16 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 17 [-1, 4] 1 0 models.common.Concat [1]18 -1 1 90880 models.common.C3 [256, 128, 1, False] 19 -1 1 147712 models.common.Conv [128, 128, 3, 2] 20 [-1, 15] 1 0 models.common.Concat [1]21 -1 1 296448 models.common.C3 [256, 256, 1, False] 22 -1 1 590336 models.common.Conv [256, 256, 3, 2] 23 [-1, 11] 1 0 models.common.Concat [1]24 -1 1 1182720 models.common.C3 [512, 512, 1, False] 25 [18, 21, 24] 1 16182 models.yolo.Detect [1, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]]YOLOv5sEMA summary: 222 layers, 7063542 parameters, 7063542 gradients, 16.2 GFLOPs参考
https://www.bilibili.com/video/BV1s84y1775U/?spm_id_from=333.788&vd_source=f83457e2adc10b543ae4c742fba1e3b2
https://blog.csdn.net/hhhhhhhhhhwwwwwwwwww/article/details/131347981
来源地址：https://blog.csdn.net/qq_43615485/article/details/131470922
