快速网页开发_河北省建设厅_哈尔滨优化调整人员流动管理_免费产品推广软件

• 🍨 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

一、实验目的：

1. 了解并学习InceptionV3相对于InceptionV1改进了哪些地方（重点）
2. 使用Inception v3完成天气识别案例

二、实验环境：

• 语言环境：python 3.8
• 编译器：Jupyter notebook
• 深度学习环境：Pytorch
  • torch==2.4.0+cu124
  • torchvision==0.19.0+cu124

三、Inception v3

Inception v3论文：Rethinking the Inception Architecture for Computer Vision

Inception v3是Inception网络系列的第三个版本，它在ImageNet图像识别竞赛中取得了优异成绩，尤其是在大规模图像识别任务中表现出色。

Inception v3的主要特点如下：

1. 更深的网络结构：Inception v3比之前的Inception网络结构更深，包含了48层卷积层。这使得网络可以提取更多层次的特征，从而在图像识别任务上取得更好的效果。
2. 使用Factorized Convolutions：Inception v3采用了Factorized Convolutions（分解卷积），将较大的卷积核分解为多个较小的卷积核。这种方法可以降低网络的参数数量，减少计算复杂度，同时保持良好的性能。
3. 使用Batch Normalization：Inception v3在每个卷积层之后都添加了Batch Normalization（BN），这有助于网络的收敛和泛化能力。BN可以减少Internal Covariate Shift（内部协变量偏移）现象，加快训练速度，同时提高模型的鲁棒性。
4. 辅助分类器：Inception v3引入了辅助分类器，可以在网络训练过程中提供额外的梯度信息，帮助网络更好地学习特征。辅助分类器位于网络的某个中间层，其输出会与主分类器的输出进行加权融合，从而得到最终的预测结果。
5. 基于RMSProp的优化器：Inception v3使用了RMSProp优化器进行训练。相比于传统的随机梯度下降（SGD）方法，RMSProp可以自适应地调整学习率，使得训练过程更加稳定，收敛速度更快。

Inception v3在图像分类、物体检测和图像分割等计算机视觉任务中均取得了显著的效果。然而，由于其较大的网络结构和计算复杂度，Inception v3在实际应用中可能需要较高的硬件要求。

相对于Inception v1的Inception Module结构，Inception v3中做出了如下改动：

将 5×5 的卷积分解为两个 3×3 的卷积运算以提升计算速度。尽管这有点违反直觉，但一个 5×5 的卷积在计算成本上是一个 3×3 卷积的 2.78 倍。所以叠加两个 3×3 卷积实际上在性能上会有所提升，如下图所示：

此外，作者将 n×n 的卷积核尺寸分解为 1×n 和 n×1 两个卷积。例如，一个 3×3 的卷积等价于首先执行一个 1×3 的卷积再执行一个 3×1 的卷积。他们还发现这种方法在成本上要比单个 3×3 的卷积降低 33%，这一结构如下图所示：

此处如果 n=3，则与上一张图像一致。最左侧的 5x5 卷积可被表示为两个 3x3 卷积，它们又可以被表示为 1x3 和 3x1 卷积。
模块中的滤波器组被扩展（即变得更宽而不是更深），以解决表征性瓶颈。如果该模块没有被拓展宽度，而是变得更深，那么维度会过多减少，造成信息损失。如下图所示：

最后实现的inception v3网络是上图结构图如下：

四、Pytorch复现

设置GPU、导入数据、划分数据集等步骤同前。

1. 构建模型

class BasicConv2d(nn.Module):def __init__(self, in_channel, out_channel, **kwargs):super(BasicConv2d, self).__init__()self.conv = nn.Conv2d(in_channel, out_channel, bias=False, **kwargs)self.norm = nn.BatchNorm2d(out_channel, eps=0.001)self.relu = nn.ReLU(inplace=True)def forward(self, x):x = self.conv(x)x = self.norm(x)x = self.relu(x)return x

class InceptionA(nn.Module):def __init__(self, in_channels, pool_features):super(InceptionA, self).__init__()self.branch1x1 = BasicConv2d(in_channels, 64, kernel_size=1) # 1self.branch5x5_1 = BasicConv2d(in_channels, 48, kernel_size=1)self.branch5x5_2 = BasicConv2d(48, 64, kernel_size=5, padding=2)self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, padding=1)self.branch_pool = BasicConv2d(in_channels, pool_features, kernel_size=1)def forward(self, x):branch1x1 = self.branch1x1(x)branch5x5 = self.branch5x5_1(x)branch5x5 = self.branch5x5_2(branch5x5)branch3x3dbl = self.branch3x3dbl_1(x)branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)branch_pool = self.branch_pool(branch_pool)outputs = [branch1x1, branch5x5, branch3x3dbl, branch_pool]return torch.cat(outputs, 1)class InceptionB(nn.Module):def __init__(self, in_channels, channels_7x7):super(InceptionB, self).__init__()self.branch1x1 = BasicConv2d(in_channels, 192, kernel_size=1)c7 = channels_7x7self.branch7x7_1 = BasicConv2d(in_channels, c7, kernel_size=1)self.branch7x7_2 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))self.branch7x7_3 = BasicConv2d(c7, 192, kernel_size=(7, 1), padding=(3, 0))self.branch7x7dbl_1 = BasicConv2d(in_channels, c7, kernel_size=1)self.branch7x7dbl_2 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))self.branch7x7dbl_3 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))self.branch7x7dbl_4 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))self.branch7x7dbl_5 = BasicConv2d(c7, 192, kernel_size=(1, 7), padding=(0, 3))self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)def forward(self, x):branch1x1 = self.branch1x1(x)branch7x7 = self.branch7x7_1(x)branch7x7 = self.branch7x7_2(branch7x7)branch7x7 = self.branch7x7_3(branch7x7)branch7x7dbl = self.branch7x7dbl_1(x)branch7x7dbl = self.branch7x7dbl_2(branch7x7dbl)branch7x7dbl = self.branch7x7dbl_3(branch7x7dbl)branch7x7dbl = self.branch7x7dbl_4(branch7x7dbl)branch7x7dbl = self.branch7x7dbl_5(branch7x7dbl)branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)branch_pool = self.branch_pool(branch_pool)outputs = [branch1x1, branch7x7, branch7x7dbl, branch_pool]return torch.cat(outputs, 1)class InceptionC(nn.Module):def __init__(self, in_channels):super(InceptionC, self).__init__()self.branch1x1 = BasicConv2d(in_channels, 320, kernel_size=1)self.branch3x3_1 = BasicConv2d(in_channels, 384, kernel_size=1)self.branch3x3_2a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))self.branch3x3_2b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))self.branch3x3dbl_1 = BasicConv2d(in_channels, 448, kernel_size=1)self.branch3x3dbl_2 = BasicConv2d(448, 384, kernel_size=3, padding=1)self.branch3x3dbl_3a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))self.branch3x3dbl_3b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)def forward(self, x):branch1x1 = self.branch1x1(x)branch3x3 = self.branch3x3_1(x)branch3x3 = [self.branch3x3_2a(branch3x3),self.branch3x3_2b(branch3x3), ]branch3x3 = torch.cat(branch3x3, 1) branch3x3dbl = self.branch3x3dbl_1(x) branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl) branch3x3dbl = [ self.branch3x3dbl_3a(branch3x3dbl), self.branch3x3dbl_3b(branch3x3dbl), ]branch3x3dbl = torch.cat(branch3x3dbl, 1) branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)branch_pool = self.branch_pool(branch_pool) outputs = [branch1x1, branch3x3, branch3x3dbl, branch_pool]return torch.cat(outputs, 1)  

class ReductionA(nn.Module):def __init__(self, in_channels):super(ReductionA, self).__init__()self.branch3x3 = BasicConv2d(in_channels, 384, kernel_size=3, stride=2)self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, stride=2)def forward(self, x):branch3x3 = self.branch3x3(x)branch3x3dbl = self.branch3x3dbl_1(x)branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)outputs = [branch3x3, branch3x3dbl, branch_pool]return torch.cat(outputs, 1)class ReductionB(nn.Module):def __init__(self, in_channels):super(ReductionB, self).__init__()self.branch3x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)self.branch3x3_2 = BasicConv2d(192, 320, kernel_size=3, stride=2)self.branch7x7x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)self.branch7x7x3_2 = BasicConv2d(192, 192, kernel_size=(1, 7), padding=(0, 3))self.branch7x7x3_3 = BasicConv2d(192, 192, kernel_size=(7, 1), padding=(3, 0))self.branch7x7x3_4 = BasicConv2d(192, 192, kernel_size=3, stride=2)def forward(self, x):branch3x3 = self.branch3x3_1(x)branch3x3 = self.branch3x3_2(branch3x3)branch7x7x3 = self.branch7x7x3_1(x)branch7x7x3 = self.branch7x7x3_2(branch7x7x3)branch7x7x3 = self.branch7x7x3_3(branch7x7x3)branch7x7x3 = self.branch7x7x3_4(branch7x7x3)branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)outputs = [branch3x3, branch7x7x3, branch_pool]return torch.cat(outputs, 1)

class InceptionAux(nn.Module):def __init__(self, in_channels, num_classes):super(InceptionAux, self).__init__()self.conv0 = BasicConv2d(in_channels, 128, kernel_size=1)self.conv1 = BasicConv2d(128, 768, kernel_size=5)self.conv1.stddev = 0.01self.fc = nn.Linear(768, num_classes)self.fc.stddev = 0.001def forward(self, x):# 17 x 17 x 768x = F.avg_pool2d(x, kernel_size=5, stride=3)# 5 x 5 x 768x = self.conv0(x)# 5 x 5 x 128x = self.conv1(x)# 1 x 1 x 768x = x.view(x.size(0), -1)# 768x = self.fc(x)# 1000return x

import torch.nn.functional as F
class InceptionV3(nn.Module):def __init__(self, num_classes=1000, aux_logits=False, transform_input=False):super(InceptionV3, self).__init__()self.aux_logits = aux_logitsself.transform_input = transform_inputself.Conv2d_1a_3x3 = BasicConv2d(3, 32, kernel_size=3, stride=2)self.Conv2d_2a_3x3 = BasicConv2d(32, 32, kernel_size=3)self.Conv2d_2b_3x3 = BasicConv2d(32, 64, kernel_size=3, padding=1)self.Conv2d_3b_1x1 = BasicConv2d(64, 80, kernel_size=1)self.Conv2d_4a_3x3 = BasicConv2d(80, 192, kernel_size=3)self.Mixed_5b = InceptionA(192, pool_features=32)self.Mixed_5c = InceptionA(256, pool_features=64)self.Mixed_5d = InceptionA(288, pool_features=64)self.Mixed_6a = ReductionA(288)self.Mixed_6b = InceptionB(768, channels_7x7=128)self.Mixed_6c = InceptionB(768, channels_7x7=160)self.Mixed_6d = InceptionB(768, channels_7x7=160)self.Mixed_6e = InceptionB(768, channels_7x7=192)if aux_logits:self.AuxLogits = InceptionAux(768, num_classes)self.Mixed_7a = ReductionB(768)self.Mixed_7b = InceptionC(1280)self.Mixed_7c = InceptionC(2048)self.fc = nn.Linear(2048, num_classes)def forward(self, x):if self.transform_input: # 1x = x.clone()x[:, 0] = x[:, 0] * (0.229 / 0.5) + (0.485 - 0.5) / 0.5x[:, 1] = x[:, 1] * (0.224 / 0.5) + (0.456 - 0.5) / 0.5x[:, 2] = x[:, 2] * (0.225 / 0.5) + (0.406 - 0.5) / 0.5# 299 x 299 x 3x = self.Conv2d_1a_3x3(x)# 149 x 149 x 32x = self.Conv2d_2a_3x3(x)# 147 x 147 x 32x = self.Conv2d_2b_3x3(x)# 147 x 147 x 64x = F.max_pool2d(x, kernel_size=3, stride=2)# 73 x 73 x 64x = self.Conv2d_3b_1x1(x)# 73 x 73 x 80x = self.Conv2d_4a_3x3(x)# 71 x 71 x 192x = F.max_pool2d(x, kernel_size=3, stride=2)# 35 x 35 x 192x = self.Mixed_5b(x)# 35 x 35 x 256x = self.Mixed_5c(x)# 35 x 35 x 288x = self.Mixed_5d(x)# 35 x 35 x 288x = self.Mixed_6a(x)# 17 x 17 x 768x = self.Mixed_6b(x)# 17 x 17 x 768x = self.Mixed_6c(x)# 17 x 17 x 768x = self.Mixed_6d(x)# 17 x 17 x 768x = self.Mixed_6e(x)# 17 x 17 x 768if self.training and self.aux_logits:aux = self.AuxLogits(x)# 17 x 17 x 768x = self.Mixed_7a(x)# 8 x 8 x 1280x = self.Mixed_7b(x)# 8 x 8 x 2048x = self.Mixed_7c(x)# 8 x 8 x 2048x = F.avg_pool2d(x, kernel_size=5)# 1 x 1 x 2048x = F.dropout(x, training=self.training)# 1 x 1 x 2048x = x.view(x.size(0), -1)# 2048x = self.fc(x)# 1000 (num_classes)if self.training and self.aux_logits:return x, auxreturn x

# 统计模型参数量以及其他指标
import torchsummary# 调用并将模型转移到GPU中
model = InceptionV3().to(device)# 显示网络结构
torchsummary.summary(model, (3, 299, 299))
print(model)

代码输出部分截图：

2. 编写训练与测试函数

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

# 编写测试函数
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losstarget_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 /= sizetest_loss /= num_batchesreturn test_acc, test_loss

3. 设置损失函数和学习率

import copyloss_fn = nn.CrossEntropyLoss()
learn_rate = 1e-4
opt = torch.optim.Adam(model.parameters(), lr=learn_rate)scheduler = torch.optim.lr_scheduler.StepLR(opt, step_size=1, gamma=0.9)  # 定义学习率高度器epochs = 100  # 设置训练模型的最大轮数为100，但可能到不了100
patience = 10  # 早停的耐心值，即如果模型连续10个周期没有准确率提升，则跳出训练train_loss = []
train_acc = []
test_loss = []
test_acc = []
best_acc = 0  # 设置一个最佳的准确率，作为最佳模型的判别指标
no_improve_epoch = 0  # 用于跟踪准确率是否提升的计数器
epoch = 0  # 用于统计最终的训练模型的轮数，这里设置初始值为0；为绘图作准备，这里的绘图范围不是epochs = 100

4. 正式训练

# 开始训练
for epoch in range(epochs):model.train()epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, opt)model.eval()epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)if epoch_test_acc > best_acc:best_acc = epoch_test_accbest_model = copy.deepcopy(model)no_improve_epoch = 0  # 重置计数器# 保存最佳模型的检查点PATH = 'J8_best_model.pth'torch.save({'epoch': epoch,'model_state_dict': best_model.state_dict(),'optimizer_state_dict': opt.state_dict(),'loss': epoch_test_loss,}, PATH)else:no_improve_epoch += 1if no_improve_epoch >= patience:print(f"Early stop triggered at epoch {epoch + 1}")break  # 早停train_acc.append(epoch_train_acc)train_loss.append(epoch_train_loss)test_acc.append(epoch_test_acc)test_loss.append(epoch_test_loss)scheduler.step()  # 更新学习率lr = opt.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))

代码输出部分截图：

5. 结果可视化

# 结果可视化
# 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(epoch)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()

6. 预测

from PIL import Imageclasses = 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)  # 展示预测的图片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'预测结果是：{pred_class}')import os
from pathlib import Path
import random#从所有的图片的随机选择一张图片image=[]
def image_path(data_dir):file_list=os.listdir(data_dir)                       #列出四个分类标签data_file_dir=file_list                              #从四个分类标签中随机选择一个data_dir=Path(data_dir)for i in data_file_dir:i=Path(i)image_file_path=data_dir.joinpath(i)            #拼接路径data_file_paths=image_file_path.iterdir()       #罗列文件夹的内容data_file_paths=list(data_file_paths)           #要转换为列表image.append(data_file_paths)file=random.choice(image)                           #从所有的图像中随机选择一类file=random.choice(file)                            #从选择的类中随机选择一张图片return filedata_dir='./weather_photos'
image_path=image_path(data_dir)# 预测训练集中的某张照片
predict_one_image(image_path=image_path,model=model,transform=train_transforms,classes=classes)

总结

Inception v1和v3都是Google的深度学习研究团队开发的深度卷积神经网络架构，它们在图像识别任务上取得了显著的成绩。以下是对两个模型的分析和它们的主要改进之处：

将5×5卷积分解为两个3×3卷积：

! []（https：//i-blog.csdnimg.cn/direct/13c3dda8d7694c788fd792a814dc7232.png）

将 n×n 的卷积核尺寸分解为 1×n 和 n×1 两个卷积 ：

Inception v1

Inception v1是2014年提出的，它的主要贡献在于引入了Inception模块，这种模块可以并行地执行不同尺度的卷积操作，从而在不同的尺度上捕捉图像特征。v1的特点包括：

• 1x1卷积核：用于降维和减少计算量，同时保持了网络的深度，有助于提取更复杂的特征。
• 不同尺度的卷积核：1x1、3x3和5x5卷积核的组合，以及3x3的最大池化，使得网络能够捕捉不同尺度的特征。
• 辅助分类器：在网络的中间层添加辅助分类器，可以帮助梯度流动，提高训练的稳定性。

Inception v3

Inception v3在v1的基础上做了许多改进，包括：

• 卷积的分解：将5x5和7x7的卷积分解成更小的卷积核，例如7x7卷积分解成两个3x3卷积，这样可以减少参数数量和计算量，同时保持网络性能。
• 辅助分类器的改进：v3中的辅助分类器在训练过程中并没有加速收敛，但提高了最终的准确率，尤其是在训练后期。
• 标签平滑正则化（LSR）：在训练中使用标签平滑正则化可以提高模型的泛化能力。
• 更高效的下采样：v3提出了更高效的下采样方法，通过在应用池化操作之前增加卷积层的维度，避免了表征瓶颈。

分析

J9与J8的输出结果对比可以大致看出，Inception v1在训练初期的准确率较低，但随着训练的进行，准确率逐渐提升。v1的训练损失和测试损失都相对较高，这可能意味着模型在训练过程中对数据的拟合程度不够高，或者模型的泛化能力有待提高。

而Inception v3的初始测试准确率较低，但随着训练的进行，准确率迅速提升，并且训练损失和测试损失都显著低于v1，这表明v3在训练过程中对数据的拟合程度更高，泛化能力也更强。

结论

Inception v3相比于v1，在网络结构、训练策略和正则化方法上都做了改进，这些改进使得v3在图像识别任务上的性能得到了显著提升。v3的网络结构更加高效，训练策略更加先进，正则化方法也更加有效，这些都是v3性能提升的关键因素。