作業：day43的時候我們安排大家對自己找的數據集用簡單cnn訓練，現在可以嘗試下借助這幾天的知識來實現精度的進一步提高

import torch
import torch.nn as nn
import torch.nn.functional as F
import torchvision
import torchvision.transforms as transforms
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image
import os# 設置隨機種子確保結果可復現
torch.manual_seed(42)
np.random.seed(42)# --- MODIFICATION 1: 更改為你的數據集路徑 ---
# 將 'path/to/your_dataset' 替換為你的數據集所在的根目錄
# Using a placeholder path. Please update this to your actual dataset path.
data_dir = r'/data/wangjinjun/learn/data/10 Big Cats of the Wild - Image Classification'
train_dir = os.path.join(data_dir, 'train')
test_dir = os.path.join(data_dir, 'test')# 檢查路徑是否存在
if not os.path.isdir(data_dir):print(f"Warning: Dataset directory not found at '{data_dir}'. "f"Please update the 'data_dir' variable to your dataset's path. "f"The script will fail if the path is not correct.")# Create dummy directories to allow the script to run without FileNotFoundError# You should replace this with your actual data.os.makedirs(train_dir, exist_ok=True)os.makedirs(test_dir, exist_ok=True)# Create dummy class folders and imagesfor d in [train_dir, test_dir]:for c in ['AFRICAN LEOPARD', 'TIGER']: # Example classesos.makedirs(os.path.join(d, c), exist_ok=True)Image.new('RGB', (100, 100)).save(os.path.join(d, c, 'dummy.jpg'))# --- MODIFICATION 2: 使用ImageFolder加載自定義數據集 ---
# 定義數據預處理步驟
transform = transforms.Compose([transforms.Resize((32, 32)),transforms.ToTensor(),transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])# 加載訓練集和測試集
trainset = torchvision.datasets.ImageFolder(root=train_dir, transform=transform)
testset = torchvision.datasets.ImageFolder(root=test_dir, transform=transform)# 從訓練數據集中自動獲取類別名稱和數量
classes = trainset.classes
num_classes = len(classes)
print(f"從數據集中找到 {num_classes} 個類別: {classes}")# 創建數據加載器
trainloader = torch.utils.data.DataLoader(trainset, batch_size=64, shuffle=True, num_workers=2)
testloader = torch.utils.data.DataLoader(testset, batch_size=64, shuffle=False, num_workers=2)# --- NEW: CBAM (Convolutional Block Attention Module) 實現 ---
class ChannelAttention(nn.Module):def __init__(self, in_planes, ratio=16):super(ChannelAttention, self).__init__()self.avg_pool = nn.AdaptiveAvgPool2d(1)self.max_pool = nn.AdaptiveMaxPool2d(1)self.fc = nn.Sequential(nn.Conv2d(in_planes, in_planes // ratio, 1, bias=False),nn.ReLU(),nn.Conv2d(in_planes // ratio, in_planes, 1, bias=False))self.sigmoid = nn.Sigmoid()def forward(self, x):avg_out = self.fc(self.avg_pool(x))max_out = self.fc(self.max_pool(x))out = avg_out + max_outreturn self.sigmoid(out)class SpatialAttention(nn.Module):def __init__(self, kernel_size=7):super(SpatialAttention, self).__init__()assert kernel_size in (3, 7), 'kernel size must be 3 or 7'padding = 3 if kernel_size == 7 else 1self.conv1 = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=False)self.sigmoid = nn.Sigmoid()def forward(self, x):avg_out = torch.mean(x, dim=1, keepdim=True)max_out, _ = torch.max(x, dim=1, keepdim=True)x = torch.cat([avg_out, max_out], dim=1)x = self.conv1(x)return self.sigmoid(x)class CBAM(nn.Module):def __init__(self, in_planes, ratio=16, kernel_size=7):super(CBAM, self).__init__()self.ca = ChannelAttention(in_planes, ratio)self.sa = SpatialAttention(kernel_size)def forward(self, x):x = self.ca(x) * xx = self.sa(x) * xreturn x# --- MODIFICATION 3: 動態調整CNN模型以適應你的數據集并集成CBAM ---
class SimpleCNN_CBAM(nn.Module):def __init__(self, num_classes):super(SimpleCNN_CBAM, self).__init__()self.conv1 = nn.Conv2d(3, 32, kernel_size=3, padding=1)self.cbam1 = CBAM(32)self.conv2 = nn.Conv2d(32, 64, kernel_size=3, padding=1)self.cbam2 = CBAM(64)self.conv3 = nn.Conv2d(64, 128, kernel_size=3, padding=1)self.cbam3 = CBAM(128)self.pool = nn.MaxPool2d(2, 2)# 輸入特征數128 * 4 * 4取決于輸入圖像大小和網絡結構。# 由于我們將所有圖像調整為32x32，經過3次2x2的池化后，尺寸變為 32 -> 16 -> 8 -> 4。所以這里是4*4。self.fc1 = nn.Linear(128 * 4 * 4, 512)# **重要**: 輸出層的大小現在由num_classes決定self.fc2 = nn.Linear(512, num_classes)def forward(self, x):x = F.relu(self.conv1(x))x = self.cbam1(x)x = self.pool(x)x = F.relu(self.conv2(x))x = self.cbam2(x)x = self.pool(x)x = F.relu(self.conv3(x))x = self.cbam3(x)x = self.pool(x)x = x.view(-1, 128 * 4 * 4)x = F.relu(self.fc1(x))x = self.fc2(x)return x# 初始化模型，傳入你的數據集的類別數量
model = SimpleCNN_CBAM(num_classes=num_classes)
print("帶有CBAM模塊的模型已創建")# 如果有GPU則使用GPU
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model = model.to(device)# 訓練模型函數 (現在使用傳入的trainloader)
def train_model(model, trainloader, epochs=5):criterion = nn.CrossEntropyLoss()optimizer = torch.optim.Adam(model.parameters(), lr=0.001)print("開始訓練...")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 + 1}, {i + 1:5d}] 損失: {running_loss / 100:.3f}')running_loss = 0.0print("訓練完成")# 定義模型保存路徑
model_save_path = 'my_custom_cnn_cbam.pth'# 嘗試加載預訓練模型
try:model.load_state_dict(torch.load(model_save_path, map_location=device))print(f"已從 '{model_save_path}' 加載預訓練模型")
except FileNotFoundError:print("無法加載預訓練模型，將開始訓練新模型。")train_model(model, trainloader, epochs=5) # 訓練新模型torch.save(model.state_dict(), model_save_path) # 保存訓練好的模型print(f"新模型已訓練并保存至 '{model_save_path}'")# 設置模型為評估模式
model.eval()# Grad-CAM實現 (這部分無需修改)
class GradCAM:def __init__(self, model, target_layer):self.model = modelself.target_layer = target_layerself.gradients = Noneself.activations = Noneself.register_hooks()def register_hooks(self):def forward_hook(module, input, output):self.activations = output.detach()def backward_hook(module, grad_input, grad_output):self.gradients = grad_output[0].detach()self.target_layer.register_forward_hook(forward_hook)self.target_layer.register_backward_hook(backward_hook)def generate_cam(self, input_image, target_class=None):model_output = self.model(input_image)if target_class is None:target_class = torch.argmax(model_output, dim=1).item()self.model.zero_grad()one_hot = torch.zeros_like(model_output)one_hot[0, target_class] = 1model_output.backward(gradient=one_hot, retain_graph=True)gradients = self.gradientsactivations = self.activationsweights = torch.mean(gradients, dim=(2, 3), keepdim=True)cam = torch.sum(weights * activations, dim=1, keepdim=True)cam = F.relu(cam)cam = F.interpolate(cam, size=(32, 32), mode='bilinear', align_corners=False)cam = cam - cam.min()cam = cam / cam.max() if cam.max() > 0 else camreturn cam.cpu().squeeze().numpy(), target_class# --- 示例：對測試集中的一張圖片使用Grad-CAM ---
# 初始化Grad-CAM，目標層是最后一個卷積層
# 注意：我們將 Grad-CAM 目標層更改為 self.cbam3，以可視化注意力模塊后的特征圖
grad_cam = GradCAM(model, model.cbam3)# 從測試集中獲取一張圖片
if len(testset) > 0:img, label = testset[0]img_tensor = img.unsqueeze(0).to(device)# 生成CAMcam, predicted_class_idx = grad_cam.generate_cam(img_tensor)# 可視化結果def visualize_cam(img, cam, predicted_class, true_class):img = img.cpu().permute(1, 2, 0).numpy() # 轉換回 (H, W, C)# 反歸一化以便顯示img = img * 0.5 + 0.5img = np.clip(img, 0, 1)heatmap = plt.cm.jet(cam)heatmap = heatmap[:, :, :3] # 去掉alpha通道overlay = heatmap * 0.4 + img * 0.6plt.figure(figsize=(10, 5))plt.subplot(1, 3, 1)plt.imshow(img)plt.title(f'Original Image\nTrue: {true_class}')plt.axis('off')plt.subplot(1, 3, 2)plt.imshow(heatmap)plt.title('Grad-CAM Heatmap')plt.axis('off')plt.subplot(1, 3, 3)plt.imshow(overlay)plt.title(f'Overlay\nPredicted: {predicted_class}')plt.axis('off')plt.show()# 顯示結果predicted_class_name = classes[predicted_class_idx]true_class_name = classes[label]visualize_cam(img, cam, predicted_class_name, true_class_name)
else:print("Test set is empty. Cannot perform visualization.")