모델 비교 (Model Comparison)

# 1. 라이브러리 임포트 from sklearn.model_selection import train_test_split from sklearn.metrics import accuracy_score from tensorflow.keras.models import Sequential, load_model from tensorflow.keras.layers import Dense, Flatten, Conv2D, MaxPooling2D from tensorflow.keras.optimizers import Adam from tensorflow.keras.utils import to_categorical from tensorflow.keras.applications import ResNet50 from tensorflow.keras.models import Model from tensorflow.keras.layers import GlobalAveragePooling2D, Input import numpy as np import matplotlib.pyplot as plt from tensorflow.keras.datasets import mnist # 2. MNIST 데이터셋 로드 및 전처리 (x_train, y_train), (x_test, y_test) = mnist.load_data() x_train = np.expand_dims(x_train, -1) x_test = np.expand_dims(x_test, -1) x_train = np.repeat(x_train, 3, axis=-1) # 채널을 3으로 확장 x_test = np.repeat(x_test, 3, axis=-1) x_train = x_train.astype('float32') / 255.0 x_test = x_test.astype('float32') / 255.0 y_train = to_categorical(y_train, 10) y_test = to_categorical(y_test, 10) # 데이터셋 분할 (훈련:검증 = 8:2) x_train, x_val, y_train, y_val = train_test_split(x_train, y_train, test_size=0.2, random_state=42) # 3. 기본 CNN 모델 정의 및 훈련 cnn_model = Sequential([ Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 3)), MaxPooling2D((2, 2)), Flatten(), Dense(128, activation='relu'), Dense(10, activation='softmax') ]) cnn_model.compile(optimizer=Adam(), loss='categorical_crossentropy', metrics=['accuracy']) cnn_history = cnn_model.fit(x_train, y_train, epochs=10, batch_size=32, validation_data=(x_val, y_val)) cnn_model.save('cnn_model.h5') # 4. ResNet50 모델 정의 및 파인튜닝 input_tensor = Input(shape=(28, 28, 3)) base_model = ResNet50(include_top=False, weights='imagenet', input_tensor=input_tensor) x = GlobalAveragePooling2D()(base_model.output) x = Dense(10, activation='softmax')(x) resnet_model = Model(inputs=base_model.input, outputs=x) for layer in base_model.layers: layer.trainable = False # ResNet50의 가중치는 고정 resnet_model.compile(optimizer=Adam(), loss='categorical_crossentropy', metrics=['accuracy']) resnet_history = resnet_model.fit(x_train, y_train, epochs=10, batch_size=32, validation_data=(x_val, y_val)) resnet_model.save('resnet_model.h5') # 5. 두 모델의 성능 비교 cnn_loss, cnn_accuracy = cnn_model.evaluate(x_test, y_test) resnet_loss, resnet_accuracy = resnet_model.evaluate(x_test, y_test) print(f"CNN Model - Test loss: {cnn_loss}, Test accuracy: {cnn_accuracy}") print(f"ResNet Model - Test loss: {resnet_loss}, Test accuracy: {resnet_accuracy}") # 그래프 그리기 plt.figure(figsize=(12, 5)) # 정확도 그래프 plt.subplot(1, 2, 1) plt.plot(cnn_history.history['accuracy'], label='CNN Train Accuracy') plt.plot(cnn_history.history['val_accuracy'], label='CNN Val Accuracy') plt.plot(resnet_history.history['accuracy'], label='ResNet Train Accuracy') plt.plot(resnet_history.history['val_accuracy'], label='ResNet Val Accuracy') plt.title('Model Accuracy') plt.xlabel('Epochs') plt.ylabel('Accuracy') plt.legend() # 손실 그래프 plt.subplot(1, 2, 2) plt.plot(cnn_history.history['loss'], label='CNN Train Loss') plt.plot(cnn_history.history['val_loss'], label='CNN Val Loss') plt.plot(resnet_history.history['loss'], label='ResNet Train Loss') plt.plot(resnet_history.history['val_loss'], label='ResNet Val Loss') plt.title('Model Loss') plt.xlabel('Epochs') plt.ylabel('Loss') plt.legend() plt.show()

import torch import torch.nn as nn import torch.optim as optim from torchvision import datasets, transforms, models from torch.utils.data import DataLoader # 데이터셋 로드 및 전처리 transform = transforms.Compose([ transforms.Resize((224, 224)), transforms.Grayscale(num_output_channels=3), # 흑백 이미지를 3채널로 변환 transforms.ToTensor(), transforms.Normalize((0.5,), (0.5,)) ]) train_dataset = datasets.MNIST(root='./data', train=True, download=True, transform=transform) test_dataset = datasets.MNIST(root='./data', train=False, download=True, transform=transform) train_loader = DataLoader(train_dataset, batch_size=32, shuffle=True) test_loader = DataLoader(test_dataset, batch_size=32, shuffle=False) # VGG16 모델 구성 및 학습 vgg16 = models.vgg16(pretrained=True) vgg16.classifier[6] = nn.Linear(4096, 10) # 출력 클래스 수정 vgg16 = vgg16.cuda() criterion = nn.CrossEntropyLoss() optimizer = optim.Adam(vgg16.parameters(), lr=0.001) # VGG16 모델 학습 함수 def train_model(model, criterion, optimizer, train_loader, num_epochs=1): # 에포크 수를 1로 줄임 model.train() for epoch in range(num_epochs): running_loss = 0.0 for i, (images, labels) in enumerate(train_loader): images, labels = images.cuda(), labels.cuda() optimizer.zero_grad() outputs = model(images) loss = criterion(outputs, labels) loss.backward() optimizer.step() running_loss += loss.item() * images.size(0) # 중간 로그 출력 if (i + 1) % 100 == 0: print(f'Batch {i+1}/{len(train_loader)}, Loss: {loss.item():.4f}') epoch_loss = running_loss / len(train_loader.dataset) print(f'Epoch {epoch+1}/{num_epochs}, Loss: {epoch_loss:.4f}') train_model(vgg16, criterion, optimizer, train_loader) # ResNet50 모델 구성 및 학습 resnet50 = models.resnet50(pretrained=True) resnet50.fc = nn.Linear(resnet50.fc.in_features, 10) # 출력 클래스 수정 resnet50 = resnet50.cuda() optimizer = optim.Adam(resnet50.parameters(), lr=0.001) train_model(resnet50, criterion, optimizer, train_loader) # 모델 평가 함수 def evaluate_model(model, test_loader): model.eval() correct = 0 total = 0 with torch.no_grad(): for images, labels in test_loader: images, labels = images.cuda(), labels.cuda() outputs = model(images) _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum().item() accuracy = 100 * correct / total return accuracy vgg16_accuracy = evaluate_model(vgg16, test_loader) resnet50_accuracy = evaluate_model(resnet50, test_loader) print(f'VGG16 Accuracy: {vgg16_accuracy:.2f}%') print(f'ResNet50 Accuracy: {resnet50_accuracy:.2f}%')