kaggle找到一個圖像數據集，用cnn網絡進行訓練并且用grad-cam做可視化

進階：并拆分成多個文件

fruit_cnn_project/
├─ data/                # 存放數據集（需手動創建，后續放入圖片）
│  ├─ train/            # 訓練集圖像
│  └─ val/              # 驗證集圖像
├─ models/              # 模型定義
│  └─ cnn_model.py      # CNN網絡結構
├─ utils/               # 工具函數
│  ├─ dataset_utils.py  # 數據加載與預處理
│  ├─ grad_cam.py       # Grad-CAM可視化
│  └─ train_utils.py    # 訓練與評估
├─ main.py              # 主程序
└─ requirements.txt     # 依賴列表（可選）

# 第一部分：導入庫
import os
import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms
import torch.nn.functional as F
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline# 第二部分：數據加載與預處理
def load_data():data_transform = transforms.Compose([transforms.Resize((224, 224)),transforms.ToTensor(),transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])train_dataset = datasets.ImageFolder(root='data/train', transform=data_transform)train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=32, shuffle=True)test_dataset = datasets.ImageFolder(root='data/test', transform=data_transform)test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=32, shuffle=False)return train_loader, test_loader# 第三部分：模型定義
class SimpleCNN(nn.Module):def __init__(self):super(SimpleCNN, self).__init__()self.conv1 = nn.Conv2d(3, 16, kernel_size=3, padding=1)self.relu = nn.ReLU()self.pool = nn.MaxPool2d(2, 2)self.conv2 = nn.Conv2d(16, 32, kernel_size=3, padding=1)self.fc1 = nn.Linear(32 * 56 * 56, 128)self.fc2 = nn.Linear(128, 2)def forward(self, x):x = self.pool(self.relu(self.conv1(x)))x = self.pool(self.relu(self.conv2(x)))x = x.view(-1, 32 * 56 * 56)x = self.relu(self.fc1(x))x = self.fc2(x)return x# 第四部分：模型訓練
train_loader, _ = load_data()
model = SimpleCNN()
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.001)
num_epochs = 10for epoch in range(num_epochs):running_loss = 0.0for i, data in enumerate(train_loader, 0):inputs, labels = dataoptimizer.zero_grad()outputs = model(inputs)loss = criterion(outputs, labels)loss.backward()optimizer.step()running_loss += loss.item()print(f'Epoch {epoch + 1}, Loss: {running_loss / len(train_loader)}')torch.save(model.state_dict(), 'trained_model.pth')# 第五部分：模型測試
_, test_loader = load_data()
model = SimpleCNN()
model.load_state_dict(torch.load('trained_model.pth'))
model.eval()
correct = 0
total = 0with torch.no_grad():for data in test_loader:images, labels = dataoutputs = model(images)_, predicted = torch.max(outputs.data, 1)total += labels.size(0)correct += (predicted == labels).sum().item()print(f'Accuracy of the network on the test images: {100 * correct / total}%')# 第六部分：Grad-CAM可視化（修復版）
def get_activation():activation = {}def hook(model, input, output):activation['target_layer'] = output.detach()return hook, activationdef grad_cam(model, image, target_class_index):hook, activation = get_activation()target_layer = model.conv2target_layer.register_forward_hook(hook)model.eval()image = image.unsqueeze(0)image.requires_grad_(True)output = model(image)one_hot = torch.zeros(1, output.size()[-1]).to(image.device)one_hot[0][target_class_index] = 1output.backward(gradient=one_hot, retain_graph=True)gradients = image.grad[0].cpu().numpy()# 從activation字典中獲取激活圖activation_map = activation['target_layer'].cpu().numpy()[0]weights = np.mean(gradients, axis=(1, 2))cam = np.zeros(activation_map.shape[1:], dtype=np.float32)for i, w in enumerate(weights):cam += w * activation_map[i]cam = np.maximum(cam, 0)cam = F.interpolate(torch.from_numpy(cam).unsqueeze(0).unsqueeze(0), size=(224, 224), mode='bilinear', align_corners=False)[0][0].numpy()cam = (cam - cam.min()) / (cam.max() - cam.min())return cam# 可視化前幾張測試圖片
dataiter = iter(test_loader)
images, labels = dataiter.next()for i in range(5):  # 可視化前5張圖片image = images[i]label = labels[i].item()cam = grad_cam(model, image, label)plt.figure(figsize=(10, 5))plt.subplot(1, 2, 1)plt.imshow(image.permute(1, 2, 0).numpy())plt.title(f'Original Image (Class: {label})')plt.axis('off')plt.subplot(1, 2, 2)plt.imshow(image.permute(1, 2, 0).numpy())plt.imshow(cam, cmap='jet', alpha=0.5)plt.title('Grad-CAM Visualization')plt.axis('off')plt.tight_layout()plt.show()