【YOLO改进】主干插入A2Attention模块(基于MMYOLO)

A2Attention模块

论文链接: https://arxiv.org/pdf/1810.11579.pdf

将A2Attention模块添加到MMYOLO中

1. 将开源代码A2Attention.py文件复制到mmyolo/models/plugins目录下

2. 导入MMYOLO用于注册模块的包: from mmyolo.registry import MODELS

3. 确保class DoubleAttention中的输入维度数名称为in_channels(因为MMYOLO会提前传入输入维度参数，所以要保持参数名的一致)

4. 利用@MODELS.register_module()将“class DoubleAttention(nn.Module)”注册:

5. 修改mmyolo/models/plugins/__init__.py文件

6. 在终端运行:

   python setup.py install

7. 修改对应的配置文件，并且将plugins的参数“type”设置为“DoubleAttention”，可参考【YOLO改进】主干插入注意力机制模块CBAM（基于MMYOLO）-优快云博客

修改后的A2Attention.py

import torch
from torch import nn
from torch.nn import functional as F
from torch.nn import init
from mmyolo.registry import MODELS


@MODELS.register_module()
class DoubleAttention(nn.Module):

    def __init__(self, in_channels,c_m=128,c_n=128,reconstruct = True):
        super().__init__()
        self.in_channels=in_channels
        self.reconstruct = reconstruct
        self.c_m=c_m
        self.c_n=c_n
        self.convA=nn.Conv2d(in_channels,c_m,1)
        self.convB=nn.Conv2d(in_channels,c_n,1)
        self.convV=nn.Conv2d(in_channels,c_n,1)
        if self.reconstruct:
            self.conv_reconstruct = nn.Conv2d(c_m, in_channels, kernel_size = 1)
        self.init_weights()


    def init_weights(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):
        b, c, h,w=x.shape
        assert c==self.in_channels
        A=self.convA(x) #b,c_m,h,w
        B=self.convB(x) #b,c_n,h,w
        V=self.convV(x) #b,c_n,h,w
        tmpA=A.view(b,self.c_m,-1)
        attention_maps=F.softmax(B.view(b,self.c_n,-1))
        attention_vectors=F.softmax(V.view(b,self.c_n,-1