365天深度学习训练营-第P8周-YOLOv5-C3模块实现

🍨 本文为🔗365天深度学习训练营 中的学习记录博客
🍖 原作者：K同学啊|接辅导、项目定制

---

任务：

• YOLO系列文章
• 思考：
  • C3模块的作用
  • C3模块为什么要这么设计？

一、模块详解

YOLOv5 C3

1.ConvBNSiLU模块

• SiLU激活函数：Sigmoid Linear Unit
  
  • 特点： 相对于ReLU函数，SiLU函数在接近零时具有更平滑的曲线，并且由于其使用了sigmoid函数，可以使网络的输出范围在0和1之间
  • 注意点：在使用SiLU时，如果数据存在过大或过小的情况，可能会导致梯度消失或梯度爆炸，因此需要进行一些调整，例如对输入数据进行归一化等。而ReLU在这方面较为稳定，不需要过多的处理。

代码解析：

class Conv(nn.Module):
    # Standard convolution
    def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True):
        # ch_in, ch_out, kernel, stride, padding, groups
        super().__init__()
        self.conv = nn.Conv2d(c1, c2, k, s, autopad(k, p), groups=g, bias=False)
        self.bn = nn.BatchNorm2d(c2)
        self.act = nn.SiLU() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())

1）自动Padding

（参考知乎某博主，见引用。）

def autopad(k, p=None): # kernel, padding
    # Pad to 'same'
    if p is None:
        p = k // 2 if isinstance(k, int) else [x // 2 for x in k] # auto-pad
    return p

为什么要自动Padding？且为什么是除以2？

• 答：
  • 一方面是偷懒，方便模块化，第二，是为了实现图片特征图经过一个以上的标准卷积之后，特征图的大小保持不变。
  • 参考上图知乎博主的推导，简单理解是Padding一般是在图片上下左右都做0填充，所以是双倍的，所以要除以2，见下图的公式。
    

2）Conv2d参数groups

• 通道分组的参数，输入通道数、输出通道数必须同时满足被groups整除
• 如果group=1，就等于是6个大小为6KK的卷积核，另外，多通道卷积的过程如下，原博客参看卷积过程详细讲解

3）pytorch nn.Identity()

pytorch源码

其实就是类似下面的代码：

def f(x):
	return x

• 不区分参数的占位符标识运算符:就是这个网络层的设计是用于占位的，即不干活，只是有这么一个层，放到残差网络里就是在跳过连接的地方用这个层，显得没有那么空虚！

2.Bottleneck模块

代码如下：

class Bottleneck(nn.Module):
    # Standard bottleneck
    def __init__(self, c1, c2, shortcut=True, g=1, e=0.5):
        # ch_in, ch_out, shortcut, groups, expansion
        super().__init__()
        c_ = int(c2 * e) # hidden channels
        self.cv1 = Conv(c1, c_, 1, 1)
        self.cv2 = Conv(c_, c2, 3, 1, g=g)
        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))

• 由两个标准卷积和一个残差连接组成
• 其中，输入输出的卷积通道数相等

二、代码

1.数据集

• 数据来源（天气识别）

data_dir = pathlib.Path("./data/")
data_paths = list(data_dir.glob('*'))

classNames = [str(path).split('/')[1] for path in data_paths]

from PIL import Image
cloudy = list(data_dir.glob('cloudy/*.jpg'))
PIL.Image.open(cloudy[100])

划分数据集

from torchvision import transforms, datasets
train_transforms = transforms.Compose([
    transforms.Resize([224, 224]),
    transforms.ToTensor(),
    transforms.Normalize(
        mean=[0.485, .456, 0.406],
        std=[0.229, 0.224, 0.225]
    )
])


total_data = datasets.ImageFolder("./data/", transform=train_transforms)
total_data

train_size = int(0.8 * len(total_data))
test_size = len(total_data) - train_size
train_dataset, test_dataset = torch.utils.data.random_split(total_data, [train_size, test_size])
train_dataset, test_dataset

batch_size = 32

train_dl = torch.utils.data.DataLoader(train_dataset, 
                                       batch_size=batch_size, 
                                       shuffle=True, 
                                       num_workers=1)

test_dl = torch.utils.data.DataLoader(test_dataset, 
                                      batch_size=batch_size, 
                                      shuffle=False,
                                      num_workers=1)

---

2.搭建模型

YOLOv5 C3 模块

import torch.nn as nn

def autopad(k, p=None): # kernel, padding
    # Pad to 'same'
    if p is None:
        p = k // 2 if isinstance(k, int) else [x // 2 for x in k] # auto-pad
    return p

class Conv(nn.Module):
    # Standard convolution
    def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True):
        # ch_in, ch_out, kernel, stride, padding, groups
        super().__init__()
        self.conv = nn.Conv2d(c1, c2, k, s, autopad(k, p), groups=g, bias=False)
        self.bn = nn.BatchNorm2d(c2)
        self.act = nn.SiLU() if act is True else (act if isinstance(act, nn.Module) else nn.Identity())
        
    def forward(self, x):
        return self.act(self.bn(self.conv(x)))
    
class Bottleneck(nn.Module):
    # Standard bottleneck
    def __init__(self, c1, c2, shortcut=True, g=1, e=0.5):
        # ch_in, ch_out, shortcut, groups, expansion
        super().__init__()
        c_ = int(c2 * e) # hidden channels
        self.cv1 = Conv(c1, c_, 1, 1)
        self.cv2 = Conv(c_, c2, 3, 1, g=g)
        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))
    
class C3(nn.Module):
    # CSP Bottleneck with 3 convolutions
    def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5):
        # ch_in, ch_out, number, shorcut, groups, expansion
        super().__init__()
        c_ = int(c2 * e) # hidden channels
        self.cv1 = Conv(c1, c_, 1, 1)
        self.cv2 = Conv(c1, c_, 1, 1, g=g)
        self.cv3 = Conv(2 * c_, c2, 1) # act = FReLU(c2)
        self.main = nn.Sequential(*(Bottleneck(c_, c_, shortcut, g, e=1.0) for _ in range(n)))
        
    def forward(self, x):
        return self.cv3(torch.cat((self.main(self.cv1(x)), self.cv2(x)), dim=1))

网络结构

• c3 + 全连接

class MyModel(nn.Module):
    def __init__(self):
        super(MyModel, self).__init__()
        
        # Conv
        self.Conv = Conv(3, 32, 3, 2)
        
        # C3
        self.c3_1 = C3(32, 64, 3, 2)
        
        # fulled-connacted
        self.classifier = nn.Sequential(
            nn.Linear(in_features=802816, out_features=100),
            nn.ReLU(),
            nn.Linear(in_features=100, out_features=len(classNames))
        )
        
    def forward(self, x):
        x = self.Conv(x)
        x = self.c3_1(x)
        x = torch.flatten(x, start_dim=1)
        x = self.classifier(x)
        
        return x

---

3.训练模型

设置训练设备

device = 'cuda' if torch.cuda.is_available() else 'cpu'
print("Using {} device".format(device))

初始化模型

model = MyModel().to(device)
model

模型参数量

import torchsummary  as summary
summary.summary(model, (3, 224, 224))

训练测试模块

训练

def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)
    num_batches = len(dataloader)
    
    train_loss, train_acc = 0, 0
    
    for X, y in dataloader:
        X, y = X.to(device), y.to(device)
        
        # loss
        pred = model(X)
        loss = loss_fn(pred, y)
        
        # back
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        
        # log
        train_loss += loss.item()
        train_acc += (pred.argmax(1) == y).type(torch.float).sum().item()
        
    train_acc /= size
    train_loss /= num_batches
    
    return train_acc, train_loss

测试

def test(dataloader, model, loss_fn):
    size = len(dataloader.dataset)
    num_batches = len(dataloader)
    
    test_loss, test_acc = 0, 0
    
    with torch.no_grad():
        for imgs, target in dataloader:
            imgs, target = imgs.to(device), target.to(device)
            
            # loss
            target_pred = model(imgs)
            loss = loss_fn(target_pred, target)
            
            test_loss += loss.item()
            test_acc += (target_pred.argmax(1) == target).type(torch.float).sum().item()
            
    test_acc /= size
    test_loss /= num_batches
    
    return test_acc, test_loss

正式训练

import copy

loss_fn = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters(), lr=1e-4)
epochs = 20

train_loss, train_acc = [], []
test_loss, test_acc = [], []

best_acc = 0

for epoch in range(epochs):
    
    model.train()
    epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)
    # scheduler.step()
    
    model.eval()
    epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
    
    if epoch_test_acc > best_acc:
        best_acc = epoch_test_acc
        best_model = copy.deepcopy(model)
        
    train_acc.append(epoch_train_acc)
    train_loss.append(epoch_train_loss)
    test_acc.append(epoch_test_acc)
    test_loss.append(epoch_test_loss)
    
    lr = optimizer.state_dict()['param_groups'][0]['lr']
    
    template = "Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%, Test_loss:{:.3f}, Lr:{:.2E}"
    print(template.format(epoch+1, 
                          epoch_train_acc*100, epoch_train_loss, 
                          epoch_test_acc*100, epoch_test_loss, 
                          lr))
    
PATH = './best_model.pth'
torch.save(model.state_dict(), PATH)

print('Done!!!')

• 尝试了，增加C3模块的数量，发现准确度下降，可能是模型过于复杂

---

4. 结果可视化

import matplotlib.pyplot as plt
import warnings
warnings.filterwarnings("ignore")
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False
plt.rcParams['figure.dpi'] = 100

epochs_range = range(epochs)

plt.figure(figsize=(12, 3))

plt.subplot(1, 2, 1)
plt.plot(epochs_range, train_acc, label='Train Acc.')
plt.plot(epochs_range, test_acc, label='Test Acc.')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')

plt.subplot(1, 2, 2)
plt.plot(epochs_range, train_loss, label='Train Loss')
plt.plot(epochs_range, test_loss, label='Test Loss')
plt.legend(loc='lower right')
plt.title('Training and Validation Loss')
plt.show()

---

5.预测

from PIL import Image

classes = list(total_data.class_to_idx)

def predict_one_image(image_path, model, transform, classes):
    
    test_img = Image.open(image_path).convert('RGB')
    plt.imshow(test_img)
    
    test_img = transform(test_img)
    img = test_img.to(device).unsqueeze(0)
    
    model.eval()
    output = model(img)
    
    _, pred = torch.max(output, 1)
    pred_class = classes[pred]
    print(f'Predict is: {pred_class}')

predict_one_image(image_path='./data/sunrise/sunrise13.jpg', 
                  model=model,
                  transform=train_transforms,
                  classes=classes
                 )

---

总结：

• SiLU激活函数
• 标准卷积模块
• 遗留问题：
  • c3模块的作用
  • c3模块为什么要这样设计？

参考：
yolov5组件笔记
激活函数ReLU和SiLU的区别
图解卷积层stride，padding，kernel_size 和卷积前后特征图尺寸之间的关系
pytorch的函数中的group参数的作用
卷积过程详细讲解