Darknet53网络结构图及代码实现

Darknet53网络结构图及代码实现

本文链接： https://blog.csdn.net/leiduifan6944/article/details/104857968


Darknet是最经典的一个深层网络，结合Resnet的特点在保证对特征进行超强表达的同时又避免了网络过深带来的梯度问题，主要有Darknet19和Darknet53，当然，如果你觉得这还不够深，在你条件允许的情况下你也可以延伸到99,199,999，…。
 

1、结构图大致如下

（这张图是从网上扒来的，凑合着，懒得自己画了）

 

2、清楚结构之后，那么，这么深的网络，一层一层的写出来的话，每层还有卷积、归一化、激活…那得写多少行才能写完呀？

莫捉急，清楚了结构，再把他狠狠封装成块，一波带走

 

1、先把卷积装一块

class Conv(nn.Module):
	def __init__(self, c_in, c_out, k, s, p, bias=True):
		"""
		自定义一个卷积块，一次性完成卷积+归一化+激活，这在类似于像DarkNet53这样的深层网络编码上可以节省很多代码
		:param c_in: in_channels
		:param c_out: out_channels
		:param k: kernel_size
		:param s:  stride
		:param p: padding
		:param bias: …
		"""
		super(Conv, self).__init__()
		self.conv = nn.Sequential(
			nn.Conv2d(c_in, c_out, k, s, p),
			nn.BatchNorm2d(c_out),
			nn.LeakyReLU(0.1),
		)

	def forward(self, entry):
		return self.conv(entry)

 

2、再把残差单元装一块

class ConvResidual(nn.Module):
	def __init__(self, c_in):		# converlution * 2 + residual
		"""
		自定义残差单元，只需给出通道数，该单元完成两次卷积，并进行加残差后返回相同维度的特征图
		:param c_in: 通道数
		"""
		c = c_in // 2
		super(ConvResidual, self).__init__()
		self.conv = nn.Sequential(
			Conv(c_in, c, 1, 1, 0),		 # kernel_size = 1进行降通道
			Conv(c, c_in, 3, 1, 1),		 # 再用kernel_size = 3把通道升回去
		)

	def forward(self, entry):
		return entry + self.conv(entry)	 # 加残差，既保留原始信息，又融入了提取到的特征
# 采用 1*1 + 3*3 的形式加深网络深度，加强特征抽象

3、最后按照结构图把残差和卷积装一块

class Darknet53(nn.Module):
	def __init__(self):
		super(Darknet53, self).__init__()
		self.conv1 = Conv(3, 32, 3, 1, 1)			# 一个卷积块 = 1层卷积
		self.conv2 = Conv(32, 64, 3, 2, 1)
		self.conv3_4 = ConvResidual(64)				# 一个残差块 = 2层卷积
		self.conv5 = Conv(64, 128, 3, 2, 1)
		self.conv6_9 = nn.Sequential(				# = 4层卷积
			ConvResidual(128),
			ConvResidual(128),
		)
		self.conv10 = Conv(128, 256, 3, 2, 1)
		self.conv11_26 = nn.Sequential(				# = 16层卷积
			ConvResidual(256),
			ConvResidual(256),
			ConvResidual(256),
			ConvResidual(256),
			ConvResidual(256),
			ConvResidual(256),
			ConvResidual(256),
			ConvResidual(256),
		)
		self.conv27 = Conv(256, 512, 3, 2, 1)
		self.conv28_43 = nn.Sequential(				# = 16层卷积
			ConvResidual(512),
			ConvResidual(512),
			ConvResidual(512),
			ConvResidual(512),
			ConvResidual(512),
			ConvResidual(512),
			ConvResidual(512),
			ConvResidual(512),
		)
		self.conv44 = Conv(512, 1024, 3, 2, 1)
		self.conv45_52 = nn.Sequential(				# = 8层卷积
			ConvResidual(1024),
			ConvResidual(1024),
			ConvResidual(1024),
			ConvResidual(1024),
		)

	def forward(self, entry):
		conv1 = self.conv1(entry)
		conv2 = self.conv2(conv1)
		conv3_4 = self.conv3_4(conv2)
		conv5 = self.conv5(conv3_4)
		conv6_9 = self.conv6_9(conv5)
		conv10 = self.conv10(conv6_9)
		conv11_26 = self.conv11_26(conv10)
		conv27 = self.conv27(conv11_26)
		conv28_43 = self.conv28_43(conv27)
		conv44 = self.conv44(conv28_43)
		conv45_52 = self.conv45_52(conv44)
		return conv45_52, conv28_43, conv11_26		# YOLOv3用，所以输出了3次特征
		

 

What？名字不是叫53吗？ 这里只有52层？原因是原本的Darknet53还包括一层输出层，前52层用于特征提取，最后一层进行最终输出。这里就根据自己实际需求再定义一层或多层对前52层提取到的特征进行融合和输出。是不是感觉一下子就豁然开朗了，就53行代码尽然就把经典的深度网络模型Darknet53写出来了