第P8周：YOLOv5-C3模块实现
文为「365天深度学习训练营」内部文章
参考本文所写记录性文章，请在文章开头带上「👉声明」
本次我将利用YOLOv5算法中的C3模块搭建网络，后续理论部分介绍将在语雀以及公众号（K同学啊）中详细展开，这里主要让大家先了解C3的结构，方便后续YOLOv5算法的学习。
🏡 我的环境：
- 语言环境：Python3.12
- 编译器：VS Code
- 数据集：天气识别数据集
- 深度学习环境：PyTorch on MPS(Macbook Pro(M4 Max 128G))
- torch==2.5.1+mps
- torchvision==0.20.1
一、 前期准备
1. 设置GPU
如果设备上支持GPU就使用GPU,否则使用CPU
import torch
import torch.nn as nn
import torchvision
from torchvision import transforms, datasets
import os,PIL,pathlib,warnings
warnings.filterwarnings("ignore") #忽略警告信息
device = torch.device("mps" if torch.backends.mps.is_available() else "cpu")
device
device(type='mps')
2. 导入数据
import os,PIL,random,pathlib
data_dir = './weather_photos/'
data_dir = pathlib.Path(data_dir)
data_paths = list(data_dir.glob('*'))
classeNames = [str(path).split("/")[1] for path in data_paths]
classeNames
['cloudy', 'rain', 'shine', 'sunrise']
# 关于transforms.Compose的更多介绍可以参考：https://blog.csdn.net/qq_38251616/article/details/124878863
train_transforms = transforms.Compose([
transforms.Resize([224, 224]), # 将输入图片resize成统一尺寸
# transforms.RandomHorizontalFlip(), # 随机水平翻转
transforms.ToTensor(), # 将PIL Image或numpy.ndarray转换为tensor，并归一化到[0,1]之间
transforms.Normalize( # 标准化处理-->转换为标准正太分布（高斯分布），使模型更容易收敛
mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225]) # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])
test_transform = transforms.Compose([
transforms.Resize([224, 224]), # 将输入图片resize成统一尺寸
transforms.ToTensor(), # 将PIL Image或numpy.ndarray转换为tensor，并归一化到[0,1]之间
transforms.Normalize( # 标准化处理-->转换为标准正太分布（高斯分布），使模型更容易收敛
mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225]) # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])
total_data = datasets.ImageFolder("./weather_photos/",transform=train_transforms)
total_data
Dataset ImageFolder
Number of datapoints: 1125
Root location: ./weather_photos/
StandardTransform
Transform: Compose(
Resize(size=[224, 224], interpolation=bilinear, max_size=None, antialias=True)
ToTensor()
Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
)
total_data.class_to_idx
{'cloudy': 0, 'rain': 1, 'shine': 2, 'sunrise': 3}
3. 划分数据集
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
(<torch.utils.data.dataset.Subset at 0x17f0faa50>,
<torch.utils.data.dataset.Subset at 0x16e1dc950>)
batch_size = 4
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=True,
num_workers=1)
for X, y in test_dl:
print("Shape of X [N, C, H, W]: ", X.shape)
print("Shape of y: ", y.shape, y.dtype)
break
Shape of X [N, C, H, W]: torch.Size([4, 3, 224, 224])
Shape of y: torch.Size([4]) torch.int64
二、搭建包含C3模块的模型
1. 搭建模型
import torch.nn.functional as F
def autopad(k, p=None): # kernel, padding
# Pad to 'same'
if p is None:
# # k 是 int 整数则除以2, 若干的整数值则循环整除
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
"""
:param c1: 输入的channel值
:param c2: 输出的channel值
:param k: 卷积的kernel_size
:param s: 卷积的stride
:param p: 卷积的padding 一般是None
:param act: 激活函数类型 True就是SiLU(), False就是不使用激活函数
:param g: 卷积的groups数 =1就是普通的卷积 >1就是深度可分离卷积
"""
super().__init__()
self.conv = nn.Conv2d(c1, c2, k, s, autopad(k, p), groups=g, bias=False)
self.bn = nn.BatchNorm2d(c2)
# 若act=True, 则激活, act=False, 不激活
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
"""
:param c1: 整个Bottleneck的输入channel
:param c2: 整个Bottleneck的输出channel
:param e: expansion ratio c2*e 就是第一个卷积的输出channel=第二个卷积的输入channel
:param shortcut: bool Bottleneck中是否有shortcut，默认True
:param g: Bottleneck中的3x3卷积类型 =1普通卷积 >1深度可分离卷积
"""
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, shortcut, groups, expansion
"""
:param c1: 整个 C3 的输入channel
:param c2: 整个 C3 的输出channel
:param n: 有n个Bottleneck
:param shortcut: bool Bottleneck中是否有shortcut，默认True
:param g: C3中的3x3卷积类型 =1普通卷积 >1深度可分离卷积
:param e: expansion ratio
"""
super().__init__()
c_ = int(c2 * e) # hidden channels
self.cv1 = Conv(c1, c_, 1, 1)
self.cv2 = Conv(c1, c_, 1, 1)
self.cv3 = Conv(2 * c_, c2, 1) # act=FReLU(c2)
self.m = nn.Sequential(*(Bottleneck(c_, c_, shortcut, g, e=1.0) for _ in range(n)))
def forward(self, x):
return self.cv3(torch.cat((self.m(self.cv1(x)), self.cv2(x)), dim=1))
class model_K(nn.Module):
def __init__(self):
super(model_K, self).__init__()
# 卷积模块
self.Conv = Conv(3, 32, 3, 2)
# C3模块1
self.C3_1 = C3(32, 64, 3, 2)
# 全连接网络层，用于分类
self.classifier = nn.Sequential(
nn.Linear(in_features=802816, out_features=100),
nn.ReLU(),
nn.Linear(in_features=100, out_features=4)
)
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
print("Using {} device".format(device))
model = model_K().to(device)
model
Using mps device
model_K(
(Conv): Conv(
(conv): Conv2d(3, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
(C3_1): C3(
(cv1): Conv(
(conv): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
(cv2): Conv(
(conv): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
(cv3): Conv(
(conv): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
(m): Sequential(
(0): Bottleneck(
(cv1): Conv(
(conv): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
(cv2): Conv(
(conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
)
(1): Bottleneck(
(cv1): Conv(
(conv): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
(cv2): Conv(
(conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
)
(2): Bottleneck(
(cv1): Conv(
(conv): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
(cv2): Conv(
(conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(act): SiLU()
)
)
)
)
(classifier): Sequential(
(0): Linear(in_features=802816, out_features=100, bias=True)
(1): ReLU()
(2): Linear(in_features=100, out_features=4, bias=True)
)
)
2. 查看模型详情
# 统计模型参数量以及其他指标
# torchsummary -> torchinfo
from torchinfo import summary
model = model.to("cpu")
batch_size = 32
summary(model, input_size = (batch_size, 3, 224, 224))
===============================================================================================
Layer (type:depth-idx) Output Shape Param #
===============================================================================================
model_K [32, 4] --
├─Conv: 1-1 [32, 32, 112, 112] --
│ └─Conv2d: 2-1 [32, 32, 112, 112] 864
│ └─BatchNorm2d: 2-2 [32, 32, 112, 112] 64
│ └─SiLU: 2-3 [32, 32, 112, 112] --
├─C3: 1-2 [32, 64, 112, 112] --
│ └─Conv: 2-4 [32, 32, 112, 112] --
│ │ └─Conv2d: 3-1 [32, 32, 112, 112] 1,024
│ │ └─BatchNorm2d: 3-2 [32, 32, 112, 112] 64
│ │ └─SiLU: 3-3 [32, 32, 112, 112] --
│ └─Sequential: 2-5 [32, 32, 112, 112] --
│ │ └─Bottleneck: 3-4 [32, 32, 112, 112] 10,368
│ │ └─Bottleneck: 3-5 [32, 32, 112, 112] 10,368
│ │ └─Bottleneck: 3-6 [32, 32, 112, 112] 10,368
│ └─Conv: 2-6 [32, 32, 112, 112] --
│ │ └─Conv2d: 3-7 [32, 32, 112, 112] 1,024
│ │ └─BatchNorm2d: 3-8 [32, 32, 112, 112] 64
│ │ └─SiLU: 3-9 [32, 32, 112, 112] --
│ └─Conv: 2-7 [32, 64, 112, 112] --
│ │ └─Conv2d: 3-10 [32, 64, 112, 112] 4,096
│ │ └─BatchNorm2d: 3-11 [32, 64, 112, 112] 128
│ │ └─SiLU: 3-12 [32, 64, 112, 112] --
├─Sequential: 1-3 [32, 4] --
│ └─Linear: 2-8 [32, 100] 80,281,700
│ └─ReLU: 2-9 [32, 100] --
│ └─Linear: 2-10 [32, 4] 404
===============================================================================================
Total params: 80,320,536
Trainable params: 80,320,536
Non-trainable params: 0
Total mult-adds (Units.GIGABYTES): 17.71
===============================================================================================
Input size (MB): 19.27
Forward/backward pass size (MB): 2260.76
Params size (MB): 321.28
Estimated Total Size (MB): 2601.31
===============================================================================================
model = model.to(device)
三、 训练模型
1. 编写训练函数
# 训练循环
def train(dataloader, model, loss_fn, optimizer):
size = len(dataloader.dataset) # 训练集的大小
num_batches = len(dataloader) # 批次数目, (size/batch_size，向上取整)
train_loss, train_acc = 0, 0 # 初始化训练损失和正确率
for X, y in dataloader: # 获取图片及其标签
X, y = X.to(device), y.to(device)
# 计算预测误差
pred = model(X) # 网络输出
loss = loss_fn(pred, y) # 计算网络输出和真实值之间的差距，targets为真实值，计算二者差值即为损失
# 反向传播
optimizer.zero_grad() # grad属性归零
loss.backward() # 反向传播
optimizer.step() # 每一步自动更新
# 记录acc与loss
train_acc += (pred.argmax(1) == y).type(torch.float).sum().item()
train_loss += loss.item()
train_acc /= size
train_loss /= num_batches
return train_acc, train_loss
2. 编写测试函数
测试函数和训练函数大致相同，但是由于不进行梯度下降对网络权重进行更新，所以不需要传入优化器
def test (dataloader, model, loss_fn):
size = len(dataloader.dataset) # 测试集的大小
num_batches = len(dataloader) # 批次数目, (size/batch_size，向上取整)
test_loss, test_acc = 0, 0
# 当不进行训练时，停止梯度更新，节省计算内存消耗
with torch.no_grad():
for imgs, target in dataloader:
imgs, target = imgs.to(device), target.to(device)
# 计算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
3. 正式训练
model.train()、model.eval()训练营往期文章中有详细的介绍。
📌如果将优化器换成 SGD 会发生什么呢？请自行探索接下来发生的诡异事件的原因。
import copy
optimizer = torch.optim.Adam(model.parameters(), lr= 1e-4)
loss_fn = nn.CrossEntropyLoss() # 创建损失函数
epochs = 20
train_loss = []
train_acc = []
test_loss = []
test_acc = []
best_acc = 0 # 设置一个最佳准确率，作为最佳模型的判别指标
for epoch in range(epochs):
model.train()
epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)
model.eval()
epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
# 保存最佳模型到 best_model
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)
# 获取当前的学习率
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))
# 保存最佳模型到文件中
PATH = './best_model.pth' # 保存的参数文件名
torch.save(model.state_dict(), PATH)
print('Done')
Epoch: 1, Train_acc:70.2%, Train_loss:1.146, Test_acc:78.7%, Test_loss:0.687, Lr:1.00E-04
Epoch: 2, Train_acc:88.2%, Train_loss:0.382, Test_acc:81.3%, Test_loss:0.593, Lr:1.00E-04
Epoch: 3, Train_acc:93.2%, Train_loss:0.226, Test_acc:79.6%, Test_loss:0.695, Lr:1.00E-04
Epoch: 4, Train_acc:93.6%, Train_loss:0.179, Test_acc:88.4%, Test_loss:0.363, Lr:1.00E-04
Epoch: 5, Train_acc:98.4%, Train_loss:0.059, Test_acc:88.0%, Test_loss:0.351, Lr:1.00E-04
Epoch: 6, Train_acc:98.1%, Train_loss:0.051, Test_acc:86.2%, Test_loss:0.483, Lr:1.00E-04
Epoch: 7, Train_acc:98.7%, Train_loss:0.043, Test_acc:82.2%, Test_loss:0.763, Lr:1.00E-04
Epoch: 8, Train_acc:97.2%, Train_loss:0.123, Test_acc:88.0%, Test_loss:0.688, Lr:1.00E-04
Epoch: 9, Train_acc:97.7%, Train_loss:0.067, Test_acc:84.4%, Test_loss:0.703, Lr:1.00E-04
Epoch:10, Train_acc:98.9%, Train_loss:0.059, Test_acc:87.6%, Test_loss:0.649, Lr:1.00E-04
Epoch:11, Train_acc:98.9%, Train_loss:0.060, Test_acc:85.3%, Test_loss:0.670, Lr:1.00E-04
Epoch:12, Train_acc:99.7%, Train_loss:0.018, Test_acc:89.3%, Test_loss:0.568, Lr:1.00E-04
Epoch:13, Train_acc:99.6%, Train_loss:0.016, Test_acc:82.7%, Test_loss:1.213, Lr:1.00E-04
Epoch:14, Train_acc:95.3%, Train_loss:0.233, Test_acc:84.0%, Test_loss:1.058, Lr:1.00E-04
Epoch:15, Train_acc:98.2%, Train_loss:0.094, Test_acc:86.7%, Test_loss:0.983, Lr:1.00E-04
Epoch:16, Train_acc:98.8%, Train_loss:0.032, Test_acc:84.4%, Test_loss:1.435, Lr:1.00E-04
Epoch:17, Train_acc:97.6%, Train_loss:0.087, Test_acc:83.6%, Test_loss:1.331, Lr:1.00E-04
Epoch:18, Train_acc:98.1%, Train_loss:0.050, Test_acc:84.4%, Test_loss:1.026, Lr:1.00E-04
Epoch:19, Train_acc:99.6%, Train_loss:0.028, Test_acc:85.8%, Test_loss:1.208, Lr:1.00E-04
Epoch:20, Train_acc:99.3%, Train_loss:0.029, Test_acc:87.1%, Test_loss:1.447, Lr:1.00E-04
Done
四、 结果可视化
1. Loss与Accuracy图
import matplotlib.pyplot as plt
#隐藏警告
import warnings
warnings.filterwarnings("ignore") #忽略警告信息
plt.rcParams['font.sans-serif'] = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False # 用来正常显示负号
plt.rcParams['figure.dpi'] = 100 #分辨率
epochs_range = range(epochs)
plt.figure(figsize=(12, 3))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, train_acc, label='Training Accuracy')
plt.plot(epochs_range, test_acc, label='Test Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')
plt.subplot(1, 2, 2)
plt.plot(epochs_range, train_loss, label='Training Loss')
plt.plot(epochs_range, test_loss, label='Test Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()
2. 模型评估
best_model.eval()
epoch_test_acc, epoch_test_loss = test(test_dl, best_model, loss_fn)
epoch_test_acc, epoch_test_loss
(0.8933333333333333, 0.5680378910695996)
# 查看是否与我们记录的最高准确率一致
epoch_test_acc
0.8933333333333333
3. SGD结果
效果和 Adam 基本相当
epoch_test_acc, epoch_test_loss
(0.8888888888888888, 0.5477136437036454)
epoch_test_acc
0.8888888888888888
扫一扫
2138
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
