【YOLOv5-6.x】设置可学习权重结合BiFPN（Add操作）

前言

在之前的这篇博客中，简要介绍了BiFPN的原理，以及YOLOv5作者如何结合BiFPN：【魔改YOLOv5-6.x（中）】：加入ACON激活函数、CBAM和CA注意力机制、加权双向特征金字塔BiFPN

本文将尝试进一步结合BiFPN，主要参考自：YOLOv5结合BiFPN

修改yaml文件（以yolov5s为例）

只修改一处

本文以yolov5s.yaml为例进行修改，修改模型配置文件时要注意以下几点：

这里的yaml文件只修改了一处，也就是将19层的Concat换成了BiFPN_Add，要想修改其他层的Concat，可以类比进行修改
BiFPN_Add本质是add操作，不是concat操作，因此，BiFPN_Add的各个输入层要求大小完全一致（通道数、feature map大小等），因此，这里要修改之前的参数[-1, 13, 6]，来满足这个要求：
- -1层就是上一层的输出，原来上一层的输出channel数为256，这里改成512
- 13层就是这里[-1, 3, C3, [512, False]], # 13
- 这样修改后，BiFPN_Add各个输入大小都是[bs,256,40,40]
- 最后BiFPN_Add后面的参数层设置为[256, 256]也就是输入输出channel数都是256

# YOLOv5 🚀 by Ultralytics, GPL-3.0 license

# Parameters
nc: 80  # number of classes
depth_multiple: 0.33  # model depth multiple
width_multiple: 0.50  # layer channel multiple
anchors:
  - [10,13, 16,30, 33,23]  # P3/8
  - [30,61, 62,45, 59,119]  # P4/16
  - [116,90, 156,198, 373,326]  # P5/32

# YOLOv5 v6.0 backbone
backbone:
  # [from, number, module, args]
  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
   [-1, 3, C3, [128]],
   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
   [-1, 6, C3, [256]],
   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
   [-1, 9, C3, [512]],
   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
   [-1, 3, C3, [1024]],
   [-1, 1, SPPF, [1024, 5]],  # 9
  ]

# YOLOv5 v6.0 BiFPN head
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 13

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)

   [-1, 1, Conv, [512, 3, 2]],  # 为了BiFPN正确add，调整channel数
   [[-1, 13, 6], 1, BiFPN_Add3, [256, 256]],  # cat P4 <--- BiFPN change 注意v5s通道数是默认参数的一半
   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]],
   [[-1, 10], 1, Concat, [1]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)

   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]

将Concat全部换成BiFPN_Add

# YOLOv5 🚀 by Ultralytics, GPL-3.0 license

# Parameters
nc: 80  # number of classes
depth_multiple: 0.33  # model depth multiple
width_multiple: 0.50  # layer channel multiple
anchors:
  - [10,13, 16,30, 33,23]  # P3/8
  - [30,61, 62,45, 59,119]  # P4/16
  - [116,90, 156,198, 373,326]  # P5/32

# YOLOv5 v6.0 backbone
backbone:
  # [from, number, module, args]
  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
   [-1, 3, C3, [128]],
   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
   [-1, 6, C3, [256]],
   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
   [-1, 9, C3, [512]],
   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
   [-1, 3, C3, [1024]],
   [-1, 1, SPPF, [1024, 5]],  # 9
  ]

# YOLOv5 v6.0 BiFPN head
head:
  [[-1, 1, Conv, [512, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 6], 1, BiFPN_Add2, [256, 256]],  # cat backbone P4
   [-1, 3, C3, [512, False]],  # 13

   [-1, 1, Conv, [256, 1, 1]],
   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
   [[-1, 4], 1, BiFPN_Add2, [128, 128]],  # cat backbone P3
   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)

   [-1, 1, Conv, [512, 3, 2]],  # 为了BiFPN正确add，调整channel数
   [[-1, 13, 6], 1, BiFPN_Add3, [256, 256]],  # cat P4 <--- BiFPN change 注意v5s通道数是默认参数的一半
   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)

   [-1, 1, Conv, [512, 3, 2]],
   [[-1, 10], 1, BiFPN_Add2, [256, 256]],  # cat head P5
   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)

   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
  ]

打印模型参数

可以参考这篇博客：【YOLOv5-6.x】模型参数及detect层输出测试（自用），进行模型配置文件测试并查看输出结果：

                 from  n    params  module                                  arguments                     
  0                -1  1      3520  models.common.Conv                      [3, 32, 6, 2, 2]              
  1                -1  1     18560  models.common.Conv                      [32, 64, 3, 2]                
  2                -1  1     18816  models.common.C3                        [64, 64, 1]                   
  3                -1  1     73984  models.common.Conv                      [64, 128, 3, 2]               
  4                -1  2    115712  models.common.C3                        [128, 128, 2]                 
  5                -1  1    295424  models.common.Conv                      [128, 256, 3, 2]              
  6                -1  3    625152  models.common.C3                        [256, 256, 3]                 
  7                -1  1   1180672  models.common.Conv                      [256, 512, 3, 2]              
  8                -1  1   1182720  models.common.C3                        [512, 512, 1]                 
  9                -1  1    656896  models.common.SPPF                      [512, 512, 5]                 
 10                -1  1    131584  models.common.Conv                      [512, 256, 1, 1]              
 11                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          
 12           [-1, 6]  1     65794  models.common.BiFPN_Add2                [256, 256]                    
 13                -1  1    296448  models.common.C3                        [256, 256, 1, False]          
 14                -1  1     33024  models.common.Conv                      [256, 128, 1, 1]              
 15                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          
 16           [-1, 4]  1     16514  models.common.BiFPN_Add2                [128, 128]                    
 17                -1  1     74496  models.common.C3                        [128, 128, 1, False]          
 18                -1  1    295424  models.common.Conv                      [128, 256, 3, 2]              
 19       [-1, 13, 6]  1     65795  models.common.BiFPN_Add3                [256, 256]                    
 20                -1  1    296448  models.common.C3                        [256, 256, 1, False]          
 21                -1  1    590336  models.common.Conv                      [256, 256, 3, 2]              
 22          [-1, 10]  1     65794  models.common.BiFPN_Add2                [256, 256]                    
 23                -1  1   1051648  models.common.C3                        [256, 512, 1, False]          
 24      [17, 20, 23]  1    229245  models.yolo.Detect                      [80, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]]
Model Summary: 278 layers, 7384006 parameters, 7384006 gradients, 17.2 GFLOPs

修改common.py

复制粘贴一下代码：

# 结合BiFPN 设置可学习参数 学习不同分支的权重
# 两个分支add操作
class BiFPN_Add2(nn.Module):
    def __init__(self, c1, c2):
        super(BiFPN_Add2, self).__init__()
        # 设置可学习参数 nn.Parameter的作用是：将一个不可训练的类型Tensor转换成可以训练的类型parameter
        # 并且会向宿主模型注册该参数 成为其一部分 即model.parameters()会包含这个parameter
        # 从而在参数优化的时候可以自动一起优化
        self.w = nn.Parameter(torch.ones(2, dtype=torch.float32), requires_grad=True)
        self.epsilon = 0.0001
        self.conv = nn.Conv2d(c1, c2, kernel_size=1, stride=1, padding=0)
        self.silu = nn.SiLU()

    def forward(self, x):
        w = self.w
        weight = w / (torch.sum(w, dim=0) + self.epsilon)
        return self.conv(self.silu(weight[0] * x[0] + weight[1] * x[1]))


# 三个分支add操作
class BiFPN_Add3(nn.Module):
    def __init__(self, c1, c2):
        super(BiFPN_Add3, self).__init__()
        self.w = nn.Parameter(torch.ones(3, dtype=torch.float32), requires_grad=True)
        self.epsilon = 0.0001
        self.conv = nn.Conv2d(c1, c2, kernel_size=1, stride=1, padding=0)
        self.silu = nn.SiLU()

    def forward(self, x):
        w = self.w
        weight = w / (torch.sum(w, dim=0) + self.epsilon)  # 将权重进行归一化
        # Fast normalized fusion
        return self.conv(self.silu(weight[0] * x[0] + weight[1] * x[1] + weight[2] * x[2]))

修改yolo.py

在parse_model函数中找到elif m is Concat:语句，在其后面加上BiFPN_Add相关语句：

elif m is Concat:
    c2 = sum(ch[x] for x in f)
# 添加bifpn_add结构
elif m in [BiFPN_Add2, BiFPN_Add3]:
    c2 = max([ch[x] for x in f])

修改train.py

1. 向优化器中添加BiFPN的权重参数

将BiFPN_Add2和BiFPN_Add3函数中定义的w参数，加入g1

    g0, g1, g2 = [], [], []  # optimizer parameter groups
    for v in model.modules():
        # hasattr: 测试指定的对象是否具有给定的属性，返回一个布尔值
        if hasattr(v, 'bias') and isinstance(v.bias, nn.Parameter):  # bias
            g2.append(v.bias)  # biases
        if isinstance(v, nn.BatchNorm2d):  # weight (no decay)
            g0.append(v.weight)
        elif hasattr(v, 'weight') and isinstance(v.weight, nn.Parameter):  # weight (with decay)
            g1.append(v.weight)
        # BiFPN_Concat
        elif isinstance(v, BiFPN_Add2) and hasattr(v, 'w') and isinstance(v.w, nn.Parameter):
            g1.append(v.w)
        elif isinstance(v, BiFPN_Add3) and hasattr(v, 'w') and isinstance(v.w, nn.Parameter):
            g1.append(v.w)

2. 查看BiFPN_Add层参数更新情况

想要查看BiFPN_Add层的参数更新情况，可以参考这篇博客【Pytorch】查看模型某一层的参数数值（自用），直接定位到w参数，随着模型训练输出对应的值。

References

YOLOv5结合BiFPN

【论文笔记】EfficientDet（BiFPN）（2020）

nn.Module、nn.Sequential和torch.nn.parameter学习笔记

文章出处登录后可见！

已经登录？立即刷新