【学习笔记】深度学习实战 | LeNet

简要声明

---

1. 学习相关网址
   1. [双语字幕]吴恩达深度学习deeplearning.ai
   2. Papers With Code
   3. Datasets
2. 深度学习网络基于PyTorch学习架构，代码测试可跑。
3. 本学习笔记单纯是为了能对学到的内容有更深入的理解，如果有错误的地方，恳请包容和指正。

参考文献

---

1. PyTorch Tutorials [https://pytorch.org/tutorials/]
2. PyTorch Docs [https://pytorch.org/docs/stable/index.html]
3. LeNet (1998) [Gradient-based learning applied to document recognition]

简要介绍

---

LeNet

Dataset	MNIST
Input (feature maps)	32×32 (28×28)
CONV Layers	2
FC Layers	2
Activation	Sigmoid
Output	10

代码分析

---

函数库调用

import torch
from torch import nn
from torch.utils.data import DataLoader
from torchvision import datasets
from torchvision.transforms import ToTensor

处理数据

数据下载

# 从开放数据集中下载训练数据
train_data = datasets.MNIST(
    root="data",
    train=True,
    download=True,
    transform=ToTensor(),
)

# 从开放数据集中下载测试数据
test_data = datasets.MNIST(
    root="data",
    train=False,
    download=True,
    transform=ToTensor(),
)

print(f'Number of training examples: {len(train_data)}')
print(f'Number of testing examples: {len(test_data)}')

Number of training examples: 60000
Number of testing examples: 10000

数据加载器（可选）

batch_size = 64

# 创建数据加载器
train_dataloader = DataLoader(train_data, batch_size=batch_size)
test_dataloader = DataLoader(test_data, batch_size=batch_size)

for X, y in test_dataloader:
    print(f"Shape of X [N, C, H, W]: {X.shape}")
    print(f"Shape of y: {y.shape} {y.dtype}")
    break

Shape of X [N, C, H, W]: torch.Size([64, 1, 28, 28])
Shape of y: torch.Size([64]) torch.int64

创建模型

# 选择训练设备
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
print(f"Using {device} device")

Using cuda device

class LeNet(nn.Module):
    def __init__(self, output_dim):
        super().__init__()

        self.conv_1 = nn.Sequential(
            nn.Conv2d(1, 6, kernel_size=5, stride=1, padding=2),
            nn.Sigmoid(),
            nn.MaxPool2d(kernel_size=2, stride=2)
        )

        self.conv_2 = nn.Sequential(
            nn.Conv2d(6, 16, kernel_size=5, stride=1, padding=0),
            nn.Sigmoid(),
            nn.MaxPool2d(kernel_size=2, stride=2)
        )

        self.fc_1 = nn.Sequential(
            nn.Linear(16*5*5, 120),
            nn.Sigmoid()
        )

        self.fc_2 = nn.Sequential(
            nn.Linear(120, 84),
            nn.Sigmoid()
        )

        self.fc_3 = nn.Linear(84, output_dim)

    def forward(self, x):
        x = self.conv_1(x)
        x = self.conv_2(x)
        x = x.view(x.size(0), -1)
        x = self.fc_1(x)
        x = self.fc_2(x)
        x = self.fc_3(x)
        return x

model = LeNet(10).to(device)
print(model)

LeNet(
(conv_1): Sequential(
(0): Conv2d(1, 6, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(1): Sigmoid()
(2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
)
(conv_2): Sequential(
(0): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))
(1): Sigmoid()
(2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
)
(fc_1): Sequential(
(0): Linear(in_features=400, out_features=120, bias=True)
(1): Sigmoid()
)
(fc_2): Sequential(
(0): Linear(in_features=120, out_features=84, bias=True)
(1): Sigmoid()
)
(fc_3): Linear(in_features=84, out_features=10, bias=True)
)

训练模型

选择损失函数和优化器

loss_fn = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters(), lr=1e-3)

训练循环

def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)
    model.train()
    for batch, (X, y) in enumerate(dataloader):
        X, y = X.to(device), y.to(device)

        # Compute prediction error
        pred = model(X)
        loss = loss_fn(pred, y)

        # Backpropagation
        loss.backward()
        optimizer.step()
        optimizer.zero_grad()

        if batch % 100 == 0:
            loss, current = loss.item(), (batch + 1) * len(X)
            print(f"loss: {loss:>7f}  [{current:>5d}/{size:>5d}]")

测试循环

def test(dataloader, model, loss_fn):
    size = len(dataloader.dataset)
    num_batches = len(dataloader)
    model.eval()
    test_loss, correct = 0, 0
    with torch.no_grad():
        for X, y in dataloader:
            X, y = X.to(device), y.to(device)
            pred = model(X)
            test_loss += loss_fn(pred, y).item()
            correct += (pred.argmax(1) == y).type(torch.float).sum().item()
    test_loss /= num_batches
    correct /= size
    print(f"Test Error: \n Accuracy: {(100*correct):>0.1f}%, Avg loss: {test_loss:>8f} \n")

训练模型

epochs = 10.
for t in range(epochs):
    print(f"Epoch {t+1}\n-------------------------------")
    train(train_dataloader, model, loss_fn, optimizer)
    test(test_dataloader, model, loss_fn)
print("Done!")

Epoch 10
loss: 0.015569 [ 64/60000]
loss: 0.029817 [ 6464/60000]
loss: 0.043169 [12864/60000]
loss: 0.027709 [19264/60000]
loss: 0.021492 [25664/60000]
loss: 0.011533 [32064/60000]
loss: 0.045418 [38464/60000]
loss: 0.042875 [44864/60000]
loss: 0.152001 [51264/60000]
loss: 0.040214 [57664/60000]
Test Error:
Accuracy: 98.6%, Avg loss: 0.044844

模型处理

保存模型

model_name = 'LeNet'
model_file = model_name + ".pth"
torch.save(model.state_dict(), model_file)
print("Saved PyTorch Model State to " + model_file)

Saved PyTorch Model State to LeNet.pth

Summary

---

安装torchsummary

pip install torchsummary

调用summary

from torchsummary import summary

model = LeNet(10).to(device)
summary(model, (1, 28, 28))

----------------------------------------------------------------
        Layer (type)               Output Shape         Param #
================================================================
            Conv2d-1            [-1, 6, 28, 28]             156
           Sigmoid-2            [-1, 6, 28, 28]               0
         MaxPool2d-3            [-1, 6, 14, 14]               0
            Conv2d-4           [-1, 16, 10, 10]           2,416
           Sigmoid-5           [-1, 16, 10, 10]               0
         MaxPool2d-6             [-1, 16, 5, 5]               0
            Linear-7                  [-1, 120]          48,120
           Sigmoid-8                  [-1, 120]               0
            Linear-9                   [-1, 84]          10,164
          Sigmoid-10                   [-1, 84]               0
           Linear-11                   [-1, 10]             850
================================================================
Total params: 61,706
Trainable params: 61,706
Non-trainable params: 0
----------------------------------------------------------------
Input size (MB): 0.00
Forward/backward pass size (MB): 0.11
Params size (MB): 0.24
Estimated Total Size (MB): 0.35
----------------------------------------------------------------