YOLOv8改进——使用C2f-Faster-EMA替换C2f

睡懒觉的痒痒鼠

已于 2024-01-14 14:33:03 修改

阅读量7.3k

点赞数 24

分类专栏： YOLOv8改进文章标签： python 深度学习计算机视觉

于 2024-01-14 14:29:07 首次发布

本文链接：https://blog.youkuaiyun.com/pope888/article/details/135540525

版权

YOLOv8改进专栏收录该内容

4 篇文章

订阅专栏

YOLOv8改进

第三章使用C2f-Faster-EMA替换C2f.(C2f-Faster-EMA推荐可以放在主干上,Neck和head部分可以选择C2f-Faster)

一、yolov8-C2f-Faster-EMA模型

1.yolov8-C2f-Faster-EMA模型结构

# Ultralytics YOLO 🚀, AGPL-3.0 license
# YOLOv8 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=yolov8n.yaml' will call yolov8.yaml with scale 'n'
  # [depth, width, max_channels]
  n: [0.33, 0.25, 1024]  # YOLOv8n summary: 225 layers,  3157200 parameters,  3157184 gradients,   8.9 GFLOPs
  s: [0.33, 0.50, 1024]  # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients,  28.8 GFLOPs
  m: [0.67, 0.75, 768]   # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients,  79.3 GFLOPs
  l: [1.00, 1.00, 512]   # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs
  x: [1.00, 1.25, 512]   # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs

# YOLOv8.0n 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, 3, C2f_Faster_EMA, [128, True]]
  - [-1, 1, Conv, [256, 3, 2]]  # 3-P3/8
  - [-1, 6, C2f_Faster_EMA, [256, True]]
  - [-1, 1, Conv, [512, 3, 2]]  # 5-P4/16
  - [-1, 6, C2f_Faster_EMA, [512, True]]
  - [-1, 1, Conv, [1024, 3, 2]]  # 7-P5/32
  - [-1, 3, C2f_Faster_EMA, [1024, True]]
  - [-1, 1, SPPF, [1024, 5]]  # 9

# YOLOv8.0n head
head:
  - [-1, 1, nn.Upsample, [None, 2, 'nearest']]
  - [[-1, 6], 1, Concat, [1]]  # cat backbone P4
  - [-1, 3, C2f_Faster_EMA, [512]]  # 12

  - [-1, 1, nn.Upsample, [None, 2, 'nearest']]
  - [[-1, 4], 1, Concat, [1]]  # cat backbone P3
  - [-1, 3, C2f_Faster_EMA, [256]]  # 15 (P3/8-small)

  - [-1, 1, Conv, [256, 3, 2]]
  - [[-1, 12], 1, Concat, [1]]  # cat head P4
  - [-1, 3, C2f_Faster_EMA, [512]]  # 18 (P4/16-medium)

  - [-1, 1, Conv, [512, 3, 2]]
  - [[-1, 9], 1, Concat, [1]]  # cat head P5
  - [-1, 3, C2f_Faster_EMA, [1024]]  # 21 (P5/32-large)

  - [[15, 18, 21], 1, Detect, [nc]]  # Detect(P3, P4, P5)

2.C2f-Faster-EMA代码

class Faster_Block_EMA(nn.Module):
    def __init__(self,
                 inc,
                 dim,
                 n_div=4,
                 mlp_ratio=2,
                 drop_path=0.1,
                 layer_scale_init_value=0.0,
                 pconv_fw_type='split_cat'
                 ):
        super().__init__()
        self.dim = dim
        self.mlp_ratio = mlp_ratio
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.n_div = n_div

        mlp_hidden_dim = int(dim * mlp_ratio)

        mlp_layer = [
            Conv(dim, mlp_hidden_dim, 1),
            nn.Conv2d(mlp_hidden_dim, dim, 1, bias=False)
        ]

        self.mlp = nn.Sequential(*mlp_layer)

        self.spatial_mixing = Partial_conv3(
            dim,
            n_div,
            pconv_fw_type
        )
        self.attention = EMA(dim)
        
        self.adjust_channel = None
        if inc != dim:
            self.adjust_channel = Conv(inc, dim, 1)

        if layer_scale_init_value > 0:
            self.layer_scale = nn.Parameter(layer_scale_init_value * torch.ones((dim)), requires_grad=True)
            self.forward = self.forward_layer_scale
        else:
            self.forward = self.forward

    def forward(self, x):
        if self.adjust_channel is not None:
            x = self.adjust_channel(x)
        shortcut = x
        x = self.spatial_mixing(x)
        x = shortcut + self.attention(self.drop_path(self.mlp(x)))
        return x

    def forward_layer_scale(self, x):
        shortcut = x
        x = self.spatial_mixing(x)
        x = shortcut + self.drop_path(self.layer_scale.unsqueeze(-1).unsqueeze(-1) * self.mlp(x))
        return x

class C3_Faster_EMA(C3):
    def __init__(self, c1, c2, n=1, shortcut=False, g=1, e=0.5):
        super().__init__(c1, c2, n, shortcut, g, e)
        c_ = int(c2 * e)  # hidden channels
        self.m = nn.Sequential(*(Faster_Block_EMA(c_, c_) for _ in range(n)))

class C2f_Faster_EMA(C2f):
    def __init__(self, c1, c2, n=1, shortcut=False, g=1, e=0.5):
        super().__init__(c1, c2, n, shortcut, g, e)
        self.m = nn.ModuleList(Faster_Block_EMA(self.c, self.c) for _ in range(n))