YOLOv11融合ICCV[2023]倒残差移动块iRMB模块及相关改进思路|YOLO改进最简教程

时间:2024-11-13 14:10:01


YOLOv11v10v8使用教程:  YOLOv11入门到入土使用教程

YOLOv11改进汇总贴:YOLOv11及自研模型更新汇总 


《Rethinking Mobile Block for Efficient Attention-based Models》

一、 模块介绍

        论文链接:https://arxiv.org/pdf/2301.01146

        代码链接:https://github.com/zhangzjn/EMO

论文速览:

         倒置残差块 (IRB) 是轻量级 CNN 的基础设施,但尚未得到基于注意力的研究的认可。这项工作从统一的角度从高效的 IRB 和 Transformer 的有效组件重新思考轻量级基础设施,将基于 CNN 的 IRB 扩展到基于注意力的模型,并抽象出一个单残差 Meta Mobile Block (MMB) 用于轻量级模型设计。遵循简单但有效的设计标准,我们推导出了一个现代的倒置残差移动块 (iRMB),并构建了一个仅用 iRMB 的 ResNetlike 高效模块 (EMO) 用于下游任务。在 ImageNet-1K、COCO2017 和 ADE20K 基准测试上的广泛实验表明,我们的 EMO 优于最先进的方法,例如,EMO-1M/2M/5M 达到 71.5、75.1 和 78.4 Top-1,超过了基于等阶 CNN/注意力的模型,同时很好地权衡了参数、效率和准确性。

总结:一种结合Resnet与Transformer的模块,包含SE注意力机制。


二、 加入到YOLO中

2.1 创建脚本文件

        首先在ultralytics->nn路径下创建blocks.py脚本,用于存放模块代码。

2.2 复制代码        

        复制代码粘到刚刚创建的blocks.py脚本中,如下图所示:

class SE(nn.Module):

    def __init__(self, channel=512, reduction=16):
        super().__init__()
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.fc = nn.Sequential(
            nn.Linear(channel, channel // reduction, bias=False),
            nn.ReLU(inplace=True),
            nn.Linear(channel // reduction, channel, bias=False),
            nn.Sigmoid()
        )

    def forward(self, x):
        b, c, _, _ = x.size()
        y = self.avg_pool(x).view(b, c)
        y = self.fc(y).view(b, c, 1, 1)
        return x * y.expand_as(x)


class iRMB(nn.Module):
    def __init__(self, dim_in, dim_out, dim_head=32, norm_in=True, has_skip=True, exp_ratio=1.0,
                 act_layer='relu', v_proj=True, dw_ks=3, stride=1, dilation=1, se_ratio=0.0, window_size=7,
                 attn_s=True, attn_drop=0., drop=0., drop_path=0., v_group=False, attn_pre=False):
        super().__init__()
        self.norm = nn.BatchNorm2d(dim_in) if norm_in else nn.Identity()
        dim_mid = int(dim_in * exp_ratio)
        self.has_skip = (dim_in == dim_out and stride == 1) and has_skip
        self.attn_s = attn_s
        if self.attn_s:
            assert dim_in % dim_head == 0, 'dim should be divisible by num_heads'
            self.dim_head = dim_head
            self.window_size = window_size
            self.num_head = dim_in // dim_head
            self.scale = self.dim_head ** -0.5
            self.attn_pre = attn_pre
            self.qk = Conv(dim_in, int(dim_in * 2), k=1, act=False)
            self.v = Conv(dim_in, dim_mid, k=1, g=self.num_head if v_group else 1, act=False)
            self.attn_drop = nn.Dropout(attn_drop)
        else:
            if v_proj:
                self.v = Conv(dim_in, dim_mid, k=1, act=act_layer)
            else:
                self.v = nn.Identity()
        self.conv_local = Conv(dim_mid, dim_mid, k=dw_ks, s=stride, d=dilation, g=dim_mid,)
        self.se = SE(dim_mid, reduction=se_ratio) if se_ratio > 0.0 else nn.Identity()

        self.proj_drop = nn.Dropout(drop)
        self.proj = Conv(dim_mid, dim_out, k=1, act=False)
        self.drop_path = DropPath(drop_path) if drop_path else nn.Identity()

    def forward(self, x):
        shortcut = x
        x = self.norm(x)
        B, C, H, W = x.shape
        if self.attn_s:
            # padding
            if self.window_size <= 0:
                window_size_W, window_size_H = W, H
            else:
                window_size_W, window_size_H = self.window_size, self.window_size
            pad_l, pad_t = 0, 0
            pad_r = (window_size_W - W % window_size_W) % window_size_W
            pad_b = (window_size_H - H % window_size_H) % window_size_H
            x = F.pad(x, (pad_l, pad_r, pad_t, pad_b, 0, 0,))
            n1, n2 = (H + pad_b) // window_size_H, (W + pad_r) // window_size_W
            x = rearrange(x, 'b c (h1 n1) (w1 n2) -> (b n1 n2) c h1 w1', n1=n1, n2=n2).contiguous()
            # attention
            b, c, h, w = x.shape
            qk = self.qk(x)
            qk = rearrange(qk, 'b (qk heads dim_head) h w -> qk b heads (h w) dim_head', qk=2, heads=self.num_head,
                           dim_head=self.dim_head).contiguous()
            q, k = qk[0], qk[1]
            attn_spa = (q @ k.transpose(-2, -1)) * self.scale
            attn_spa = attn_spa.softmax(dim=-1)
            attn_spa = self.attn_drop(attn_spa)
            if self.attn_pre:
                x = rearrange(x, 'b (heads dim_head) h w -> b heads (h w) dim_head', heads=self.num_head).contiguous()
                x_spa = attn_spa @ x
                x_spa = rearrange(x_spa, 'b heads (h w) dim_head -> b (heads dim_head) h w', heads=self.num_head, h=h,
                                  w=w).contiguous()
                x_spa = self.v(x_spa)
            else:
                v = self.v(x)
                v = rearrange(v, 'b (heads dim_head) h w -> b heads (h w) dim_head', heads=self.num_head).contiguous()
                x_spa = attn_spa @ v
                x_spa = rearrange(x_spa, 'b heads (h w) dim_head -> b (heads dim_head) h w', heads=self.num_head, h=h,
                                  w=w).contiguous()
            # unpadding
            x = rearrange(x_spa, '(b n1 n2) c h1 w1 -> b c (h1 n1) (w1 n2)', n1=n1, n2=n2).contiguous()
            if pad_r > 0 or pad_b > 0:
                x = x[:, :, :H, :W].contiguous()
        else:
            x = self.v(x)

        x = x + self.se(self.conv_local(x)) if self.has_skip else self.se(self.conv_local(x))

        x = self.proj_drop(x)
        x = self.proj(x)

        x = (shortcut + self.drop_path(x)) if self.has_skip else x
        return x

2.3 更改task.py文件 

       打开ultralytics->nn->modules->task.py,在脚本空白处导入函数。

from ultralytics.nn.blocks import *

        之后找到模型解析函数parse_model(约在tasks.py脚本中940行左右位置,可能因代码版本不同变动),在该函数的最后一个else分支上面增加相关解析代码。

        elif m is iRMB:
            c2 = args[0]
            args = [ch[f], *args]

2.4 更改yaml文件 

yam文件解读:YOLO系列 “.yaml“文件解读_yolo yaml文件-****博客

       打开更改ultralytics/cfg/models/11路径下的YOLOv11.yaml文件,替换原有模块。(放在该位置仅能插入该模块,具体效果未知。博主精力有限,仅完成与其他模块二次创新融合的测试,结构图见文末,代码见群文件更新。)

# Ultralytics YOLO ????, AGPL-3.0 license
# YOLO11 object detection model with P3-P5 outputs. For Usage examples see https://docs.ultralytics.com/tasks/detect

# Parameters
nc: 80 # number of classes
scales: # model compound scaling constants, i.e. 'model=yolo11n.yaml' will call yolo11.yaml with scale 'n'
  # [depth, width, max_channels]
  n: [0.50, 0.25, 1024] # summary: 319 layers, 2624080 parameters, 2624064 gradients, 6.6 GFLOPs
  s: [0.50, 0.50, 1024] # summary: 319 layers, 9458752 parameters, 9458736 gradients, 21.7 GFLOPs
  m: [0.50, 1.00, 512] # summary: 409 layers, 20114688 parameters, 20114672 gradients, 68.5 GFLOPs
  l: [1.00, 1.00, 512] # summary: 631 layers, 25372160 parameters, 25372144 gradients, 87.6 GFLOPs
  x: [1.00, 1.50, 512] # summary: 631 layers, 56966176 parameters, 56966160 gradients, 196.0 GFLOPs

# YOLO11n 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]]
  - [-1, 1, Conv, [256, 3, 2]] # 3-P3/8
  - [-1, 2, C3k2, [512, False, 0.25]]
  - [-1, 1, Conv, [512, 3, 2]] # 5-P4/16
  - [-1, 2, iRMB, [512]]
  - [-1, 1, Conv, [1024, 3, 2]] # 7-P5/32
  - [-1, 2, C3k2, [1024, True]]
  - [-1, 1, SPPF, [1024, 5]] # 9
  - [-1, 2, C2PSA, [1024]] # 10

# YOLO11n head
head:
  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
  - [[-1, 6], 1, Concat, [1]] # cat backbone P4
  - [-1, 2, C3k2, [512, False]] # 13

  - [-1, 1, nn.Upsample, [None, 2, "nearest"]]
  - [[-1, 4], 1, Concat, [1]] # cat backbone P3
  - [-1, 2, C3k2, [256, False]] # 16 (P3/8-small)

  - [-1, 1, Conv, [256, 3, 2]]
  - [[-1, 13], 1, Concat, [1]] # cat head P4
  - [-1, 2, C3k2, [512, False]] # 19 (P4/16-medium)

  - [-1, 1, Conv, [512, 3, 2]]
  - [[-1, 10], 1, Concat, [1]] # cat head P5
  - [-1, 2, C3k2, [1024, True]] # 22 (P5/32-large)

  - [[16, 19, 22], 1, Detect, [nc]] # Detect(P3, P4, P5)


 2.5 修改train.py文件

       创建Train脚本用于训练。

from ultralytics.models import YOLO
import os
os.environ['KMP_DUPLICATE_LIB_OK'] = 'True'

if __name__ == '__main__':
    model = YOLO(model='ultralytics/cfg/models/11/yolo11.yaml')
    # model.load('yolov8n.pt')
    model.train(data='./data.yaml', epochs=2, batch=1, device='0', imgsz=640, workers=2, cache=False,
                amp=True, mosaic=False, project='runs/train', name='exp')

         在train.py脚本中填入修改好的yaml路径,运行即可训练,数据集创建教程见下方链接。

YOLOv11入门到入土使用教程(含结构图)_yolov11使用教程-****博客

三、相关改进思路(2024/11/8日群文件)

        该模块可替换C2f、C3模块中的BottleNeck部分,代码见群文件,结构如图。自研模块与该模块融合代码及yaml文件见群文件。

 ⭐另外,融合上百种深度学习改进模块的YOLO项目仅79.9(含百种改进的v9),RTDETR79.9,含高性能自研模型,更易发论文,代码每周更新,欢迎点击下方小卡片加我了解。⭐

⭐⭐平均每个文章对应4-6个二创及自研融合模块⭐⭐