YOLOv11 继承自 YOLOv8，在保持同等精度的前提下，参数量降低了约 20%。为了更直观地理解其架构演进，我们结合核心代码与结构图，逐一拆解其中的关键模块。
1. 动态卷积层
动态卷积允许网络根据输入特征自适应调整权重，提升泛化能力。
def autopad(k, p=None, d=1):
"""自动填充以保持输出形状相同"""
if d > 1:
k = d * (k - 1) + 1 if isinstance(k, int) else [d * (x - 1) + 1 for x in k]
if p is None:
p = k // 2 if isinstance(k, int) else [x // 2 for x in k]
return p
class DynamicConv(nn.Module):
"""动态卷积层"""
def __init__(self, c1, c2, k=1, s=1, p=None, g=1, d=1, act=True, num_experts=4):
super().__init__()
self.conv = nn.Sequential(
DynamicConv_Single(c1, c2, kernel_size=k, stride=s,
padding=autopad(k, p, d), dilation=d, groups=g, num_experts=num_experts),
nn.BatchNorm2d(c2),
nn.SiLU() if act else nn.Identity()
)
def forward(self, x):
return self.conv(x)
2. 标准卷积层
这是最基础的构建块，包含卷积、批归一化和激活函数。
class Conv(nn.Module):
"""卷积层，包含卷积、批归一化和激活函数"""
def __init__(self, in_channels, out_channels, kernel_size=3, stride=1, padding=None, groups=1, act=True):
super(Conv, self).__init__()
self.conv = nn.Conv2d(in_channels, out_channels, kernel_size, stride, autopad(kernel_size, padding), groups=groups, bias=False)
self.bn = nn.BatchNorm2d(out_channels)
self.act = nn.SiLU() if act else nn.Identity()
def forward(self, x):
return self.act(self.bn(self.conv(x)))
3. Bottleneck 结构
瓶颈模块通过压缩通道再扩展的方式，在减少计算量的同时保留信息。
class Bottleneck(nn.Module):
"""标准瓶颈模块，包含两个卷积层"""
def __init__(self, c1, c2, shortcut=True, g=1, k=(3, 3), e=0.5):
super(Bottleneck, self).__init__()
c_ = int(c2 * e)
self.cv1 = Conv(c1, c_, k[0], 1)
self.cv2 = Conv(c_, c2, k[1], 1)
self.add = shortcut and c1 == c2
def forward(self, x):
return x + self.cv2(self.cv1(x)) if self.add else self.cv2(self.cv1(x))
4. C2f 层
C2f 结合了多个 Bottleneck 模块，在小网络中表现优异，有助于加速处理。
class C2f(nn.Module):
def __init__(self, c1, c2, n=1, shortcut=False, g=1, e=0.5):
super().__init__()
self.c = int(c2 * e)
self.cv1 = Conv(c1, 2 * self.c, 1, 1)
self.cv2 = Conv((2 + n) * self.c, c2, 1)
self.m = nn.ModuleList(Bottleneck(self.c, self.c, shortcut, g, k=((3, 3), (3, 3)), e=1.0) for _ in range(n))
def forward(self, x):
y = list(self.cv1(x).chunk(2, 1))
y.extend(m(y[-1]) for m in self.m)
return self.cv2(torch.cat(y, 1))
def forward_split(self, x):
y = self.cv1(x).split((self.c, self.c), 1)
y = [y[0], y[1]]
y.extend(m(y[-1]) for m in self.m)
return self.cv2(torch.cat(y, 1))
5. C3k 层
C3k 是 C3 系列的改进版，通常用于平衡速度与精度。
class C3k(nn.Module):
"""C3k 模块，包含多个瓶颈模块"""
def __init__(self, c1, c2, n=1, shortcut=False, g=1, e=0.5, k=3):
super(C3k, self).__init__()
c_ = int(c2 * e)
self.m = nn.Sequential(*(Bottleneck(c1, c2, shortcut, g, k=(k, k), e=e) for _ in range(n)))
def forward(self, x):
return self.m(x)
6. C3k2 层
在大网络中，C3k2 往往替代 C3k，进一步增强了特征融合能力。
class C3k2(C2f):
def __init__(self, c1, c2, n=1, c3k=False, e=0.5, g=1, shortcut=True):
super().__init__(c1, c2, n, shortcut, g, e)
self.m = nn.ModuleList(
C3k(self.c, self.c, 2, shortcut, g) if c3k else Bottleneck(self.c, self.c, shortcut, g) for _ in range(n)
)
7. C3k 动态卷积
将动态卷积引入 C3k 结构，旨在进一步提升模型的适应性。
class C3k_DynamicConv(C3k):
"""C3k 模块，使用动态卷积"""
def __init__(self, c1, c2, n=1, shortcut=False, g=1, e=0.5, k=3):
super().__init__(c1, c2, n, shortcut, g, e, k)
c_ = int(c2 * e)
self.m = nn.Sequential(*(Bottleneck_DynamicConv(c_, c_, shortcut, g, k=(k, k), e=1.0) for _ in range(n)))
class C3k2_DynamicConv(C3k2):
"""C3k2 模块，使用动态卷积"""
def __init__(self, c1, c2, n=1, c3k=False, e=0.5, g=1, shortcut=True):
super().__init__(c1, c2, n, c3k, e, g, shortcut)
self.m = nn.ModuleList(C3k_DynamicConv(self.c, self.c, 2, shortcut, g) if c3k else Bottleneck_DynamicConv(self.c, self.c, shortcut, g, k=(3, 3), e=1.0) for _ in range(n))
8. FFN 层
前馈神经网络层负责非线性变换，增强特征表达能力。
9. PSA 层
位置敏感注意力（PSA）模块引入了自注意力机制，帮助模型关注关键区域。
class PSABlock(nn.Module):
def __init__(self, c, attn_ratio=0.5, num_heads=4, shortcut=True) -> None:
super().__init__()
self.attn = Attention(c, attn_ratio=attn_ratio, num_heads=num_heads)
self.ffn = nn.Sequential(Conv(c, c * 2, 1), Conv(c * 2, c, 1, act=False))
self.add = shortcut
def forward(self, x):
x = x + self.attn(x) if self.add else self.attn(x)
x = x + self.ffn(x) if self.add else self.ffn(x)
return x
10. C2PSA 层
结合卷积与注意力机制的混合模块，进一步优化特征提取效果。
注意力机制的卷积块：
class C2PSA(nn.Module):
def __init__(self, c1, c2, n=1, e=0.5):
super().__init__()
assert c1 == c2
self.c = int(c1 * e)
self.cv1 = Conv(c1, 2 * self.c, 1, 1)
self.cv2 = Conv(2 * self.c, c1, 1)
self.m = nn.Sequential(*(PSABlock(self.c, attn_ratio=0.5, num_heads=self.c // 64) for _ in range(n)))
def forward(self, x):
a, b = self.cv1(x).split((self.c, self.c), dim=1)
b = self.m(b)
return self.cv2(torch.cat((a, b), 1))
