💡💡💡本专栏所有程序均经过测试，可成功执行💡💡💡
本文给大家带来的教程是将YOLO26的下采样替换为DRFD来提取特征。文章在介绍主要的原理后，将手把手教学如何进行模块的代码添加和修改，并将修改后的完整代码放在文章的最后，方便大家一键运行，小白也可轻松上手实践。以帮助您更好地学习深度学习目标检测YOLO系列的挑战。
专栏地址：YOLO26改进-论文涨点——点击跳转看所有内容，关注不迷路！
目录
1.论文
2. Conv_BCN代码实现
2.1 将Conv_BCN添加到YOLO26中
2.2 更改init.py文件
2.3 添加yaml文件
2.4 在task.py中进行注册
2.5 执行程序
3. 完整代码分享
4. GFLOPs
5. 进阶
6.总结
1.论文
论文地址：A Robust Feature Downsampling Module for Remote-Sensing Visual Tasks
官方代码：官方代码仓库点击即可跳转
2. DRFD代码实现
2.1 将DRFD添加到YOLO26中
关键步骤一：在ultralytics\ultralytics\nn\modules下面新建文件夹models，在文件夹下新建DRFD.py，粘贴下面代码
import torch import torch.nn as nn class Cut(nn.Module): def __init__(self, in_channels, out_channels): super().__init__() self.conv_fusion = nn.Conv2d(in_channels * 4, out_channels, kernel_size=1, stride=1) self.batch_norm = nn.BatchNorm2d(out_channels) def forward(self, x): x0 = x[:, :, 0::2, 0::2] # x = [B, C, H/2, W/2] x1 = x[:, :, 1::2, 0::2] x2 = x[:, :, 0::2, 1::2] x3 = x[:, :, 1::2, 1::2] x = torch.cat([x0, x1, x2, x3], dim=1) # x = [B, 4*C, H/2, W/2] x = self.conv_fusion(x) # x = [B, out_channels, H/2, W/2] x = self.batch_norm(x) return x class DRFD(nn.Module): def __init__(self, in_channels, out_channels): super().__init__() self.cut_c = Cut(in_channels=in_channels, out_channels=out_channels) self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=1, padding=1, groups=in_channels) self.conv_x = nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=2, padding=1, groups=out_channels) self.act_x = nn.GELU() self.batch_norm_x = nn.BatchNorm2d(out_channels) self.batch_norm_m = nn.BatchNorm2d(out_channels) self.max_m = nn.MaxPool2d(kernel_size=2, stride=2) self.fusion = nn.Conv2d(3 * out_channels, out_channels, kernel_size=1, stride=1) def forward(self, x): # input: x = [B, C, H, W] c = x # c = [B, C, H, W] x = self.conv(x) # x = [B, C, H, W] --> [B, 2C, H, W] m = x # m = [B, 2C, H, W] # CutD c = self.cut_c(c) # c = [B, C, H, W] --> [B, 2C, H/2, W/2] # ConvD x = self.conv_x(x) # x = [B, 2C, H, W] --> [B, 2C, H/2, W/2] x = self.act_x(x) x = self.batch_norm_x(x) # MaxD m = self.max_m(m) # m = [B, 2C, H/2, W/2] m = self.batch_norm_m(m) # Concat + conv x = torch.cat([c, x, m], dim=1) # x = [B, 6C, H/2, W/2] x = self.fusion(x) # x = [B, 6C, H/2, W/2] --> [B, 2C, H/2, W/2] return x # x = [B, 2C, H/2, W/2]2.2 更改init.py文件
关键步骤二：在文件ultralytics\ultralytics\nn\modules\models文件夹下新建__init__.py文件，先导入函数
然后在下面的__all__中声明函数
2.3 添加yaml文件
关键步骤三：在/ultralytics/ultralytics/cfg/models/26下面新建文件yolo26_DRFD.yaml文件，粘贴下面的内容
- 目标检测
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license # Ultralytics YOLO26 object detection model with P3/8 - P5/32 outputs # Model docs: https://docs.ultralytics.com/models/yolo26 # Task docs: https://docs.ultralytics.com/tasks/detect # Parameters nc: 80 # number of classes end2end: True # whether to use end-to-end mode reg_max: 1 # DFL bins scales: # model compound scaling constants, i.e. 'model=yolo26n.yaml' will call yolo26.yaml with scale 'n' # [depth, width, max_channels] n: [0.50, 0.25, 1024] # summary: 260 layers, 2,572,280 parameters, 2,572,280 gradients, 6.1 GFLOPs s: [0.50, 0.50, 1024] # summary: 260 layers, 10,009,784 parameters, 10,009,784 gradients, 22.8 GFLOPs m: [0.50, 1.00, 512] # summary: 280 layers, 21,896,248 parameters, 21,896,248 gradients, 75.4 GFLOPs l: [1.00, 1.00, 512] # summary: 392 layers, 26,299,704 parameters, 26,299,704 gradients, 93.8 GFLOPs x: [1.00, 1.50, 512] # summary: 392 layers, 58,993,368 parameters, 58,993,368 gradients, 209.5 GFLOPs # YOLO26n backbone backbone: # [from, repeats, module, args] - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 - [-1, 2, C3k2, [256, False, 0.25]] # 2-P2/4 - [-1, 1, DRFD, [256]] # 3-P3/8 - [-1, 2, C3k2, [512, False, 0.25]] # 4-P3/8 - [-1, 1, DRFD, [512]] # 5-P4/16 - [-1, 2, C3k2, [512, True]] # 6-P4/16 - [-1, 1, DRFD, [1024]] # 7-P5/32 - [-1, 2, C3k2, [1024, True]] # 8-P5/32 - [-1, 1, SPPF, [1024, 5, 3, True]] # 9-P5/32 - [-1, 2, C2PSA, [1024]] # 10-P5/32 # YOLO26n head head: - [-1, 1, nn.Upsample, [None, 2, "nearest"]] # 11-P4/16 - [[-1, 6], 1, Concat, [1]] # 12-P4/16 - [-1, 2, C3k2, [512, True]] # 13-P4/16 - [-1, 1, nn.Upsample, [None, 2, "nearest"]] # 14-P3/8 - [[-1, 4], 1, Concat, [1]] # 15-P3/8 - [-1, 2, C3k2, [256, True]] # 16-P3/8 - [-1, 1, DRFD, [256]] # 17-P4/16 - [[-1, 13], 1, Concat, [1]] # 18-P4/16 - [-1, 2, C3k2, [512, True]] # 19-P4/16 - [-1, 1, DRFD, [512]] # 20-P5/32 - [[-1, 10], 1, Concat, [1]] # 21-P5/32 - [-1, 1, C3k2, [1024, True, 0.5, True]] # 22-P5/32 - [[16, 19, 22], 1, Detect, [nc]] # 23-P3/8,P4/16,P5/32- 语义分割
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license # Ultralytics YOLO26 object detection model with P3/8 - P5/32 outputs # Model docs: https://docs.ultralytics.com/models/yolo26 # Task docs: https://docs.ultralytics.com/tasks/detect # Parameters nc: 80 # number of classes end2end: True # whether to use end-to-end mode reg_max: 1 # DFL bins scales: # model compound scaling constants, i.e. 'model=yolo26n.yaml' will call yolo26.yaml with scale 'n' # [depth, width, max_channels] n: [0.50, 0.25, 1024] # summary: 260 layers, 2,572,280 parameters, 2,572,280 gradients, 6.1 GFLOPs s: [0.50, 0.50, 1024] # summary: 260 layers, 10,009,784 parameters, 10,009,784 gradients, 22.8 GFLOPs m: [0.50, 1.00, 512] # summary: 280 layers, 21,896,248 parameters, 21,896,248 gradients, 75.4 GFLOPs l: [1.00, 1.00, 512] # summary: 392 layers, 26,299,704 parameters, 26,299,704 gradients, 93.8 GFLOPs x: [1.00, 1.50, 512] # summary: 392 layers, 58,993,368 parameters, 58,993,368 gradients, 209.5 GFLOPs # YOLO26n backbone backbone: # [from, repeats, module, args] - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 - [-1, 2, C3k2, [256, False, 0.25]] # 2-P2/4 - [-1, 1, DRFD, [256]] # 3-P3/8 - [-1, 2, C3k2, [512, False, 0.25]] # 4-P3/8 - [-1, 1, DRFD, [512]] # 5-P4/16 - [-1, 2, C3k2, [512, True]] # 6-P4/16 - [-1, 1, DRFD, [1024]] # 7-P5/32 - [-1, 2, C3k2, [1024, True]] # 8-P5/32 - [-1, 1, SPPF, [1024, 5, 3, True]] # 9-P5/32 - [-1, 2, C2PSA, [1024]] # 10-P5/32 # YOLO26n head head: - [-1, 1, nn.Upsample, [None, 2, "nearest"]] # 11-P4/16 - [[-1, 6], 1, Concat, [1]] # 12-P4/16 - [-1, 2, C3k2, [512, True]] # 13-P4/16 - [-1, 1, nn.Upsample, [None, 2, "nearest"]] # 14-P3/8 - [[-1, 4], 1, Concat, [1]] # 15-P3/8 - [-1, 2, C3k2, [256, True]] # 16-P3/8 - [-1, 1, DRFD, [256]] # 17-P4/16 - [[-1, 13], 1, Concat, [1]] # 18-P4/16 - [-1, 2, C3k2, [512, True]] # 19-P4/16 - [-1, 1, DRFD, [512]] # 20-P5/32 - [[-1, 10], 1, Concat, [1]] # 21-P5/32 - [-1, 1, C3k2, [1024, True, 0.5, True]] # 22-P5/32 - [[16, 19, 22], 1, Segment, [nc, 32, 256]]- 旋转目标检测
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license # Ultralytics YOLO26 object detection model with P3/8 - P5/32 outputs # Model docs: https://docs.ultralytics.com/models/yolo26 # Task docs: https://docs.ultralytics.com/tasks/detect # Parameters nc: 80 # number of classes end2end: True # whether to use end-to-end mode reg_max: 1 # DFL bins scales: # model compound scaling constants, i.e. 'model=yolo26n.yaml' will call yolo26.yaml with scale 'n' # [depth, width, max_channels] n: [0.50, 0.25, 1024] # summary: 260 layers, 2,572,280 parameters, 2,572,280 gradients, 6.1 GFLOPs s: [0.50, 0.50, 1024] # summary: 260 layers, 10,009,784 parameters, 10,009,784 gradients, 22.8 GFLOPs m: [0.50, 1.00, 512] # summary: 280 layers, 21,896,248 parameters, 21,896,248 gradients, 75.4 GFLOPs l: [1.00, 1.00, 512] # summary: 392 layers, 26,299,704 parameters, 26,299,704 gradients, 93.8 GFLOPs x: [1.00, 1.50, 512] # summary: 392 layers, 58,993,368 parameters, 58,993,368 gradients, 209.5 GFLOPs # YOLO26n backbone backbone: # [from, repeats, module, args] - [-1, 1, Conv, [64, 3, 2]] # 0-P1/2 - [-1, 1, Conv, [128, 3, 2]] # 1-P2/4 - [-1, 2, C3k2, [256, False, 0.25]] # 2-P2/4 - [-1, 1, DRFD, [256]] # 3-P3/8 - [-1, 2, C3k2, [512, False, 0.25]] # 4-P3/8 - [-1, 1, DRFD, [512]] # 5-P4/16 - [-1, 2, C3k2, [512, True]] # 6-P4/16 - [-1, 1, DRFD, [1024]] # 7-P5/32 - [-1, 2, C3k2, [1024, True]] # 8-P5/32 - [-1, 1, SPPF, [1024, 5, 3, True]] # 9-P5/32 - [-1, 2, C2PSA, [1024]] # 10-P5/32 # YOLO26n head head: - [-1, 1, nn.Upsample, [None, 2, "nearest"]] # 11-P4/16 - [[-1, 6], 1, Concat, [1]] # 12-P4/16 - [-1, 2, C3k2, [512, True]] # 13-P4/16 - [-1, 1, nn.Upsample, [None, 2, "nearest"]] # 14-P3/8 - [[-1, 4], 1, Concat, [1]] # 15-P3/8 - [-1, 2, C3k2, [256, True]] # 16-P3/8 - [-1, 1, DRFD, [256]] # 17-P4/16 - [[-1, 13], 1, Concat, [1]] # 18-P4/16 - [-1, 2, C3k2, [512, True]] # 19-P4/16 - [-1, 1, DRFD, [512]] # 20-P5/32 - [[-1, 10], 1, Concat, [1]] # 21-P5/32 - [-1, 1, C3k2, [1024, True, 0.5, True]] # 22-P5/32 - [[16, 19, 22], 1, OBB, [nc, 1]]温馨提示：本文只是对yolo26基础上添加模块，如果要对yolo26 n/l/m/x进行添加则只需要指定对应的depth_multiple 和 width_multiple
end2end: True # whether to use end-to-end mode reg_max: 1 # DFL bins scales: # model compound scaling constants, i.e. 'model=yolo26n.yaml' will call yolo26.yaml with scale 'n' # [depth, width, max_channels] n: [0.50, 0.25, 1024] # summary: 260 layers, 2,572,280 parameters, 2,572,280 gradients, 6.1 GFLOPs s: [0.50, 0.50, 1024] # summary: 260 layers, 10,009,784 parameters, 10,009,784 gradients, 22.8 GFLOPs m: [0.50, 1.00, 512] # summary: 280 layers, 21,896,248 parameters, 21,896,248 gradients, 75.4 GFLOPs l: [1.00, 1.00, 512] # summary: 392 layers, 26,299,704 parameters, 26,299,704 gradients, 93.8 GFLOPs x: [1.00, 1.50, 512] # summary: 392 layers, 58,993,368 parameters, 58,993,368 gradients, 209.5 GFLOPs2.4 在task.py中进行注册
关键步骤四：在parse_model函数中进行注册，添加DRFD
先在task.py导入函数
然后在task.py文件下找到parse_model这个函数，如下图，添加DRFD
elif m in {DRFD}: c1, c2 = ch[f], args[0] if c2 != nc: # if c2 != nc (e.g., Classify() output) c2 = make_divisible(min(c2, max_channels) * width, 8) args = [c1, c2, *args[1:]]2.5 执行程序
关键步骤五:在ultralytics文件中新建train.py，将model的参数路径设置为yolo26_DRFD.yaml的路径即可 【注意是在外边的Ultralytics下新建train.py】
from ultralytics import YOLO import warnings warnings.filterwarnings('ignore') from pathlib import Path if __name__ == '__main__': # 加载模型 model = YOLO("ultralytics/cfg/26/yolo26.yaml") # 你要选择的模型yaml文件地址 # Use the model results = model.train(data=r"你的数据集的yaml文件地址", epochs=100, batch=16, imgsz=640, workers=4, name=Path(model.cfg).stem) # 训练模型🚀运行程序，如果出现下面的内容则说明添加成功🚀
from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 2 -1 1 6640 ultralytics.nn.modules.block.C3k2 [32, 64, 1, False, 0.25] 3 -1 1 30464 ultralytics.nn.models.DRFD.DRFD [64, 64] 4 -1 1 26080 ultralytics.nn.modules.block.C3k2 [64, 128, 1, False, 0.25] 5 -1 1 118272 ultralytics.nn.models.DRFD.DRFD [128, 128] 6 -1 1 87040 ultralytics.nn.modules.block.C3k2 [128, 128, 1, True] 7 -1 1 334848 ultralytics.nn.models.DRFD.DRFD [128, 256] 8 -1 1 346112 ultralytics.nn.modules.block.C3k2 [256, 256, 1, True] 9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5, 3, True] 10 -1 1 249728 ultralytics.nn.modules.block.C2PSA [256, 256, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 13 -1 1 119808 ultralytics.nn.modules.block.C3k2 [384, 128, 1, True] 14 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 15 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 16 -1 1 34304 ultralytics.nn.modules.block.C3k2 [256, 64, 1, True] 17 -1 1 30464 ultralytics.nn.models.DRFD.DRFD [64, 64] 18 [-1, 13] 1 0 ultralytics.nn.modules.conv.Concat [1] 19 -1 1 95232 ultralytics.nn.modules.block.C3k2 [192, 128, 1, True] 20 -1 1 118272 ultralytics.nn.models.DRFD.DRFD [128, 128] 21 [-1, 10] 1 0 ultralytics.nn.modules.conv.Concat [1] 22 -1 1 463104 ultralytics.nn.modules.block.C3k2 [384, 256, 1, True, 0.5, True] 23 [16, 19, 22] 1 309656 ultralytics.nn.modules.head.Detect [80, 1, True, [64, 128, 256]] YOLO26_DRFD summary: 295 layers, 2,539,768 parameters, 2,539,768 gradients, 6.0 GFLOPs3. 完整代码分享
主页侧边
4. GFLOPs
关于GFLOPs的计算方式可以查看：百面算法工程师 | 卷积基础知识——Convolution
未改进的YOLO26n GFLOPs
改进后的GFLOPs
5. 进阶
可以与其他的注意力机制或者损失函数等结合，进一步提升检测效果
6.总结
通过以上的改进方法，我们成功提升了模型的表现。这只是一个开始，未来还有更多优化和技术深挖的空间。在这里，我想隆重向大家推荐我的专栏——<专栏地址：YOLO26改进-论文涨点——点击跳转看所有内容，关注不迷路！>。这个专栏专注于前沿的深度学习技术，特别是目标检测领域的最新进展，不仅包含对YOLO26的深入解析和改进策略，还会定期更新来自各大顶会（如CVPR、NeurIPS等）的论文复现和实战分享。
为什么订阅我的专栏？——专栏地址：YOLO26改进-论文涨点——点击跳转看所有内容，关注不迷路！
前沿技术解读：专栏不仅限于YOLO系列的改进，还会涵盖各类主流与新兴网络的最新研究成果，帮助你紧跟技术潮流。
详尽的实践分享：所有内容实践性也极强。每次更新都会附带代码和具体的改进步骤，保证每位读者都能迅速上手。
问题互动与答疑：订阅我的专栏后，你将可以随时向我提问，获取及时的答疑。
实时更新，紧跟行业动态：不定期发布来自全球顶会的最新研究方向和复现实验报告，让你时刻走在技术前沿。
专栏适合人群：
对目标检测、YOLO系列网络有深厚兴趣的同学
希望在用YOLO算法写论文的同学
对YOLO算法感兴趣的同学等
