【YOLO改进】主干插入GC模块(基于MMYOLO)

GlobalContext（GC）模块

论文链接：https://arxiv.org/pdf/1904.11492.pdf

将GlobalContext（GC）模块添加到MMYOLO中

1. 将开源代码GC.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 GlobalContext中的输入维度为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 GlobalContext(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“GlobalContext”，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GC.py

import torch
import torch.nn.functional as F
from timm.models.layers.create_act import create_act_layer, get_act_layer
from timm.models.layers.helpers import make_divisible
from timm.models.layers.mlp import ConvMlp
from timm.models.layers.norm import LayerNorm2d
from torch import nn as nn
from mmyolo.registry import MODELS

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

    def __init__(self, in_channels, use_attn=True, fuse_add=False, fuse_scale=True, init_last_zero=False,
                 rd_ratio=1./8, rd_channels=None, rd_divisor=1, act_layer=nn.ReLU, gate_layer='sigmoid'):
        super(GlobalContext, self).__init__()
        act_layer = get_act_layer(act_layer)

        self.conv_attn = nn.Conv2d(in_channels, 1, kernel_size=1, bias=True) if use_attn else None

        if rd_channels is None:
            rd_channels = make_divisible(in_channels * rd_ratio, rd_divisor, round_limit=0.)
        if fuse_add:
            self.mlp_add = ConvMlp(in_channels, rd_channels, act_layer=act_layer, norm_layer=LayerNorm2d)
        else:
            self.mlp_add = None
        if fuse_scale:
            self.mlp_scale = ConvMlp(in_channels, rd_channels, act_layer=act_layer, norm_layer=LayerNorm2d)
        else:
            self.mlp_scale = None

        self.gate = create_act_layer(gate_layer)
        self.init_last_zero = init_last_zero
        self.reset_parameters()

    def reset_parameters(self):
        if self.conv_attn is not None:
            nn.init.kaiming_normal_(self.conv_attn.weight, mode='fan_in', nonlinearity='relu')
        if self.mlp_add is not None:
            nn.init.zeros_(self.mlp_add.fc2.weight)

    def forward(self, x):
        B, C, H, W = x.shape

        if self.conv_attn is not None:
            attn = self.conv_attn(