目錄

1、環境條件

2、代碼實現

3、總結

---

1、環境條件

1. pycharm編譯器
2. pytorch依賴
3. matplotlib依賴
4. numpy依賴等等

2、代碼實現

import torch
import torch.nn as nn
import torch.optim as optim
import torchvision
import torchvision.transforms as transforms
import matplotlib.pyplot as plt
import numpy as np# 設置設備
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")# 定義數據變換
transform = transforms.Compose([transforms.ToTensor(),transforms.Normalize((0.5,), (0.5,))
])# 加載 MNIST 數據集
trainset = torchvision.datasets.MNIST(root='./data', train=True, download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=64, shuffle=True)testset = torchvision.datasets.MNIST(root='./data', train=False, download=True, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=64, shuffle=False)# 定義 LeNet-5 模型
class LeNet5(nn.Module):def __init__(self):super(LeNet5, self).__init__()self.conv1 = nn.Conv2d(1, 6, kernel_size=5, stride=1, padding=2)self.pool = nn.MaxPool2d(kernel_size=2, stride=2)self.conv2 = nn.Conv2d(6, 16, kernel_size=5, stride=1)self.fc1 = nn.Linear(16 * 5 * 5, 120)self.fc2 = nn.Linear(120, 84)self.fc3 = nn.Linear(84, 10)def forward(self, x):x = self.pool(torch.relu(self.conv1(x)))x = self.pool(torch.relu(self.conv2(x)))x = x.view(-1, 16 * 5 * 5)x = torch.relu(self.fc1(x))x = torch.relu(self.fc2(x))x = self.fc3(x)return x# 初始化模型、損失函數和優化器
model = LeNet5().to(device)
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.001)# 訓練模型
epochs = 5
for epoch in range(epochs):running_loss = 0.0for i, data in enumerate(trainloader, 0):inputs, labels = datainputs, labels = inputs.to(device), labels.to(device)optimizer.zero_grad()outputs = model(inputs)loss = criterion(outputs, labels)loss.backward()optimizer.step()running_loss += loss.item()if i % 100 == 99:print(f'[Epoch {epoch + 1}, Batch {i + 1}] loss: {running_loss / 100:.3f}')running_loss = 0.0print('Finished Training')# 保存模型
torch.save(model.state_dict(), 'lenet5.pth')
print('Model saved to lenet5.pth')# 加載模型
model = LeNet5()
model.load_state_dict(torch.load('lenet5.pth'))
model.to(device)
model.eval()# 在測試集上評估模型
correct = 0
total = 0
with torch.no_grad():for data in testloader:images, labels = dataimages, labels = images.to(device), labels.to(device)outputs = model(images)_, predicted = torch.max(outputs.data, 1)total += labels.size(0)correct += (predicted == labels).sum().item()print(f'Accuracy on the test set: {100 * correct / total:.2f}%')# 加載并預處理本地圖片進行預測
from PIL import Imagedef load_and_preprocess_image(image_path):img = Image.open(image_path).convert('L')  # 轉為灰度圖img = img.resize((28, 28))img = np.array(img, dtype=np.float32)img = (img / 255.0 - 0.5) / 0.5  # 歸一化到[-1, 1]img = torch.tensor(img).unsqueeze(0).unsqueeze(0)  # 添加批次和通道維度return img.to(device)# 預測本地圖片
image_path = '4.png'  # 替換為你的本地圖片路徑
img = load_and_preprocess_image(image_path)# 使用加載的模型進行預測
model.eval()
with torch.no_grad():outputs = model(img)_, predicted = torch.max(outputs, 1)# 打印預測結果
predicted_label = predicted.item()
print(f'預測結果: {predicted_label}')# 顯示圖片及預測結果
img_np = img.cpu().numpy().squeeze()
plt.imshow(img_np, cmap='gray')
plt.title(f'預測結果: {predicted_label}')
plt.show()

解釋：torch.save（）方法完成模型的保存，image_path為本地圖片，用于測試

3、總結

????????安裝環境是比較難的點，均使用pip install 。。指令進行依賴環境的安裝，其他的比較簡單。

學習之所以會想睡覺，是因為那是夢開始的地方。
?(?ˊ?ˋ)??（開心） ?(?ˊ?ˋ)??（開心）?(?ˊ?ˋ)??（開心）?(?ˊ?ˋ)??（開心）?(?ˊ?ˋ)??（開心）
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ------不寫代碼不會凸的小劉