專欄介紹：YOLOv9改進系列 | 包含深度學習最新創新，主力高效漲點！！！
一、改進點介紹
????????Dynamic Snake Convolution是一種針對細長微弱的局部結構特征與復雜多變的全局形態特征設計的卷積模塊。
? ? ? ? SCConv是一種即插即用的空間和通道重建卷積。
? ? ? ? RepNCSPELAN4是YOLOv9中的特征提取模塊，類似YOLOv5和v8中的C2f與C3模塊。????????
二、DS-RepNCSPELAN4模塊詳解
?2.1 模塊簡介
???????DS-RepNCSPELAN4的主要思想：? 使用Dynamic Snake Convolution、SCConv與RepNCSPELAN4中融合。
三、?DS-RepNCSPELAN4模塊使用教程
3.1 DS-RepNCSPELAN4模塊的代碼
class RepConvN_SC(RepConvN):"""RepConv is a basic rep-style block, including training and deploy statusThis code is based on https://github.com/DingXiaoH/RepVGG/blob/main/repvgg.py"""default_act = nn.SiLU() # default activationdef __init__(self, c1, c2, k=3, s=1, p=1, g=1, d=1, act=True, bn=False, deploy=False):super().__init__(c1, c2, k, s, p, g, d, act, bn, deploy)assert k == 3 and p == 1self.g = gself.c1 = c1self.c2 = c2self.act = self.default_act if act is True else act if isinstance(act, nn.Module) else nn.Identity()self.bn = Noneself.conv1 = SCConv(c1, c2, k, s, p=p, g=g)self.conv2 = Conv(c1, c2, 1, s, p=(p - k // 2), g=g, act=False)class RepNBottleneck_SC(RepNBottleneck):# Standard bottleneckdef __init__(self, c1, c2, shortcut=True, g=1, k=(3, 3), e=0.5): # ch_in, ch_out, shortcut, kernels, groups, expandsuper().__init__( c1, c2, shortcut, g, k, e)c_ = int(c2 * e) # hidden channelsself.cv1 = RepConvN_SC(c1, c_, k[0], 1)self.cv2 = SCConv(c_, c2, k[1], s=1, g=g)self.add = shortcut and c1 == c2class RepNCSP_SCConv(RepNCSP):# CSP Bottleneck with 3 convolutionsdef __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5): # ch_in, ch_out, number, shortcut, groups, expansionsuper().__init__(c1, c2, n, shortcut, g, e)c_ = int(c2 * e) # hidden channelsself.cv1 = Conv(c1, c_)self.cv2 = Conv(c1, c_)self.cv3 = Conv(2 * c_, c2) # optional act=FReLU(c2)self.m = nn.Sequential(*(RepNBottleneck_SC(c_, c_, shortcut, g, e=1.0) for _ in range(n)))class SCConv(nn.Module):"""https://github.com/MCG-NKU/SCNet/blob/master/scnet.py"""def __init__(self, inplanes, planes,k=3, s=1, p=1, dilation=1, g=1, pooling_r=4):super(SCConv, self).__init__()self.k2 = nn.Sequential(nn.AvgPool2d(kernel_size=pooling_r, stride=pooling_r),Conv(inplanes, planes, k=k, s=s, p=p, d=dilation, g=g, act=False))self.k3 = Conv(inplanes, planes, k=k, s=s, p=p, d=dilation, g=g, act=False)self.k4 = Conv(inplanes, planes, k=k, s=s, p=p, d=dilation, g=g, act=False)def forward(self, x):identity = xout = torch.sigmoid(torch.add(identity, F.interpolate(self.k2(x), identity.size()[2:]))) # sigmoid(identity + k2)out = torch.mul(self.k3(x), out) # k3 * sigmoid(identity + k2)out = self.k4(out) # k4return outclass DS_RepNCSPELAN4(RepNCSPELAN4):# csp-elandef __init__(self, c1, c2, c3, c4, c5=1): # ch_in, ch_out, number, shortcut, groups, expansionsuper().__init__(c1, c2, c3, c4, c5)self.cv1 = Conv(c1, c3, k=1, s=1)self.cv2 = nn.Sequential(RepNCSP_SCConv(c3 // 2, c4, c5), DySnakeConv(c4, c4, 3))self.cv3 = nn.Sequential(RepNCSP_SCConv(c4, c4, c5), DySnakeConv(c4, c4, 3))self.cv4 = Conv(c3 + (2 * c4), c2, 1, 1)3.2 在YOlO v9中的添加教程
閱讀YOLOv9添加模塊教程或使用下文操作
? ? ? ? 1.?將YOLOv9工程中models下common.py文件中的最下行（否則可能因類繼承報錯）增加模塊的代碼。
?
?????????2.?將YOLOv9工程中models下yolo.py文件中的第681行（可能因版本變化而變化）增加以下代碼。
RepNCSPELAN4, SPPELAN, DS_RepNCSPELAN4}:
3.3 運行配置文件
# YOLOv9
# Powered bu https://blog.csdn.net/StopAndGoyyy# parameters
nc: 80 # number of classes
#depth_multiple: 0.33 # model depth multiple
depth_multiple: 1 # model depth multiple
#width_multiple: 0.25 # layer channel multiple
width_multiple: 1 # layer channel multiple
#activation: nn.LeakyReLU(0.1)
#activation: nn.ReLU()# anchors
anchors: 3# YOLOv9 backbone
backbone:[[-1, 1, Silence, []], # conv down[-1, 1, Conv, [64, 3, 2]], # 1-P1/2# conv down[-1, 1, Conv, [128, 3, 2]], # 2-P2/4# elan-1 block[-1, 1, DS_RepNCSPELAN4, [256, 128, 64, 1]], # 3# avg-conv down[-1, 1, ADown, [256]], # 4-P3/8# elan-2 block[-1, 1, RepNCSPELAN4, [512, 256, 128, 1]], # 5# avg-conv down[-1, 1, ADown, [512]], # 6-P4/16# elan-2 block[-1, 1, RepNCSPELAN4, [512, 512, 256, 1]], # 7# avg-conv down[-1, 1, ADown, [512]], # 8-P5/32# elan-2 block[-1, 1, RepNCSPELAN4, [512, 512, 256, 1]], # 9]# YOLOv9 head
head:[# elan-spp block[-1, 1, SPPELAN, [512, 256]], # 10# up-concat merge[-1, 1, nn.Upsample, [None, 2, 'nearest']],[[-1, 7], 1, Concat, [1]], # cat backbone P4# elan-2 block[-1, 1, RepNCSPELAN4, [512, 512, 256, 1]], # 13# up-concat merge[-1, 1, nn.Upsample, [None, 2, 'nearest']],[[-1, 5], 1, Concat, [1]], # cat backbone P3# elan-2 block[-1, 1, RepNCSPELAN4, [256, 256, 128, 1]], # 16 (P3/8-small)# avg-conv-down merge[-1, 1, ADown, [256]],[[-1, 13], 1, Concat, [1]], # cat head P4# elan-2 block[-1, 1, RepNCSPELAN4, [512, 512, 256, 1]], # 19 (P4/16-medium)# avg-conv-down merge[-1, 1, ADown, [512]],[[-1, 10], 1, Concat, [1]], # cat head P5# elan-2 block[-1, 1, RepNCSPELAN4, [512, 512, 256, 1]], # 22 (P5/32-large)# multi-level reversible auxiliary branch# routing[5, 1, CBLinear, [[256]]], # 23[7, 1, CBLinear, [[256, 512]]], # 24[9, 1, CBLinear, [[256, 512, 512]]], # 25# conv down[0, 1, Conv, [64, 3, 2]], # 26-P1/2# conv down[-1, 1, Conv, [128, 3, 2]], # 27-P2/4# elan-1 block[-1, 1, RepNCSPELAN4, [256, 128, 64, 1]], # 28# avg-conv down fuse[-1, 1, ADown, [256]], # 29-P3/8[[23, 24, 25, -1], 1, CBFuse, [[0, 0, 0]]], # 30 # elan-2 block[-1, 1, RepNCSPELAN4, [512, 256, 128, 1]], # 31# avg-conv down fuse[-1, 1, ADown, [512]], # 32-P4/16[[24, 25, -1], 1, CBFuse, [[1, 1]]], # 33 # elan-2 block[-1, 1, RepNCSPELAN4, [512, 512, 256, 1]], # 34# avg-conv down fuse[-1, 1, ADown, [512]], # 35-P5/32[[25, -1], 1, CBFuse, [[2]]], # 36# elan-2 block[-1, 1, RepNCSPELAN4, [512, 512, 256, 1]], # 37# detection head# detect[[31, 34, 37, 16, 19, 22], 1, DualDDetect, [nc]], # DualDDetect(A3, A4, A5, P3, P4, P5)]3.4 訓練過程
歡迎關注!
)
)
