[실습] CNN
# 1. 라이브러리 호출
import numpy as np
import matplotlib.pyplot as plt # pip install matplotlib
import torch
import torch.nn as nn
from torch.autograd import Variable
import torch.nn.functional as F
import torchvision
import torchvision.transforms as transforms
from torch.utils.data import Dataset,DataLoader
# 2. CPU 혹은 GPU 장치 확인
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# 3. fashion_mnist 데이터셋 내려받기
train_dataset = torchvision.datasets.FashionMNIST(r'C:\Users\032\SKT_AI', download=True, transform=transforms.Compose([transforms.ToTensor()]))
test_dataset = torchvision.datasets.FashionMNIST(r'C:\Users\032\SKT_AI', download=True, train=False, transform=transforms.Compose([transforms.ToTensor()]))
# 4. fashion_mnist 데이터를 데이터로더에 전달
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=100)
test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=100)
# 5. 분류에 사용될 클래스 정의
labels_map = {0:'T-Shirt', 1:'Trouser', 2:'Pullover', 3:'Dress', 4:'Coat', 5:'Sandal', 6:'Shirt', 7:'Sneaker', 8:'Bag', 9:'Ankle Boot'}
# 6. 심층 신경망 모델 생성성
class FashionDNN(nn.Module):
def __init__(self):
super(FashionDNN, self).__init__()
self.fc1 = nn.Linear(in_features=784, out_features=256)
self.drop = nn.Dropout(0.25)
self.fc2 = nn.Linear(in_features=256, out_features=128)
self.fc3 = nn.Linear(in_features=128, out_features=10)
def forward(self, input_data):
out = input_data.view(-1,784)
out = F.relu(self.fc1(out))
out = self.drop(out)
out = F.relu(self.fc2(out))
out = self.fc3(out)
return out
# 7. 심층 신경망에서 필요한 파라미터 정의
learning_rate = 0.001;
model = FashionDNN();
model.to(device)
criterion = nn.CrossEntropyLoss();
optimizer = torch.optim.Adam(model.parameters(), lr = learning_rate);
print(model)
# 8. 심층 신경망을 이용한 모델 학습
num_epochs = 5
count = 0
loss_list = []
iteration_list = []
accuracy_list = []
predictions_list = []
labels_list = []
for epoch in range(num_epochs):
for images, labels in train_loader:
images, labels = images.to(device), labels.to(device)
train = Variable(images.view(100,1,28,28))
labels = Variable(labels)
outputs = model(train)
loss= criterion(outputs,labels)
optimizer.zero_grad()
loss.backward()
optimizer.step()
count +=1
if not (count % 50):
total = 0
correct = 0
for images, labels in test_loader:
images,labels = images.to(device), labels.to(device)
labels_list.append(labels)
test = Variable(images.view(100,1,28,28))
outputs = model(test)
predictions = torch.max(outputs, 1)[1].to(device)
predictions_list.append(predictions)
correct += (predictions == labels).sum().item()
total += len(labels)
accuracy = correct * 100 / total
loss_list.append(loss.data)
iteration_list.append(count)
accuracy_list.append(accuracy)
if not (count%500):
print("Iteration: {}, Loss: {}, Accuracy: {}%".format(count, loss.data, accuracy))
# 9. 합성곱 네트워크 생성
class FashionCNN(nn.Module):
def __init__(self):
super(FashionCNN, self).__init__()
self.layer1 = nn.Sequential(
nn.Conv2d(in_channels=1, out_channels=32, kernel_size=3, padding=1),
nn.BatchNorm2d(32),
nn.ReLU(),
nn.MaxPool2d(kernel_size=2, stride=2)
)
self.layer2 = nn.Sequential(
nn.Conv2d(in_channels=32, out_channels=64, kernel_size=3),
nn.BatchNorm2d(64),
nn.ReLU(),
nn.MaxPool2d(2)
)
self.fc1 = nn.Linear(in_features=64*6*6, out_features=600)
self.drop = nn.Dropout2d(0.25)
self.fc2 = nn.Linear(in_features=600, out_features=120)
self.fc3 = nn.Linear(in_features=120, out_features=10)
def forward(self,x):
out = self.layer1(x)
out = self.layer2(out)
out = out.view(out.size(0),-1)
out = self.fc1(out)
out = self.drop(out)
out = self.fc2(out)
out = self.fc3(out)
return out
# 10. 합성곱 네트워크를 위한 파라미터 정의
learning_rate = 0.001;
model = FashionCNN();
model.to(device)
criterion = nn.CrossEntropyLoss();
optimizer = torch.optim.Adam(model.parameters(), lr = learning_rate);
print(model)
# 11. 모델 학습 및 평가
num_epochs = 5
count = 0
loss_list = []
iteration_list = []
accuracy_list = []
predictions_list = []
labels_list = []
for epoch in range(num_epochs):
for images, labels in train_loader:
images, labels = images.to(device), labels.to(device)
train = Variable(images.view(100,1,28,28))
labels = Variable(labels)
outputs = model(train)
loss = criterion(outputs, labels)
optimizer.zero_grad()
loss.backward()
optimizer.step()
count += 1
if not (count % 50):
total = 0
correct = 0
for images, labels in test_loader:
images, labels = images.to(device), labels.to(device)
labels_list.append(labels)
test = Variable(images.view(100,1,28,28))
outputs = model(test)
predictions = torch.ma(outputs, 1)[1].to(device)
predictions_list.append(predictions)
correct += (predictions == labels).sum()
total += len(labels)
accuracy = correct *100 / total
loss_list.append(loss.data)
iteration_list.append(count)
accuracy_list.append(accuracy)
if not (count % 500):
print("Iteration: {}, Loss:{}, Accuracy: {}%".format(count, loss.data, accuracy))
FashionDNN(
(fc1): Linear(in_features=784, out_features=256, bias=True)
(drop): Dropout(p=0.25, inplace=False)
(fc2): Linear(in_features=256, out_features=128, bias=True)
(fc3): Linear(in_features=128, out_features=10, bias=True)
)
FashionCNN(
(layer1): Sequential(
(0): Conv2d(1, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU()
(3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
)
(layer2): Sequential(
(0): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1))
(1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU()
(3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
)
(fc1): Linear(in_features=2304, out_features=600, bias=True)
(drop): Dropout2d(p=0.25, inplace=False)
(fc2): Linear(in_features=600, out_features=120, bias=True)
(fc3): Linear(in_features=120, out_features=10, bias=True)
)
c:\Users\032\anaconda3\envs\torch_book\lib\site-packages\torch\nn\functional.py:1538: UserWarning: dropout2d: Received a 2-D input to dropout2d, which is deprecated and will result in an error in a future release. To retain the behavior and silence this warning, please use dropout instead. Note that dropout2d exists to provide channel-wise dropout on inputs with 2 spatial dimensions, a channel dimension, and an optional batch dimension (i.e. 3D or 4D inputs).
warnings.warn(warn_msg)
Iteration: 500, Loss:0.5086044073104858, Accuracy: 87.44999694824219%
Iteration: 1000, Loss:0.2946288585662842, Accuracy: 88.43999481201172%
Iteration: 1500, Loss:0.262689471244812, Accuracy: 86.33999633789062%
Iteration: 2000, Loss:0.24397815763950348, Accuracy: 88.44999694824219%
Iteration: 2500, Loss:0.1428401917219162, Accuracy: 89.06999969482422%
Iteration: 3000, Loss:0.14736072719097137, Accuracy: 90.33999633789062%[실습] 전이학습
catanddog.zip
14.82MB
# 12. 라이브러리 호출
import os
import time
import copy
import glob
import cv2
import shutil
import torch
import torchvision
import torchvision.transforms as transforms
import torchvision.models as models
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import DataLoader
import matplotlib.pyplot as plt
import numpy as np
# 13. 이미지 데이터 전처리 방법 정의
data_path = r'C:\Users\032\SKT_AI\catanddog'
transform = transforms.Compose(
[
transforms.Resize([256, 256]),
transforms.RandomResizedCrop(224),
transforms.RandomHorizontalFlip(),
transforms.ToTensor()
]
)
train_dataset = torchvision.datasets.ImageFolder(
data_path,
transform=transform
)
train_loader = torch.utils.data.DataLoader(
train_dataset,
batch_size=32,
num_workers=8,
shuffle=True
)
print(len(train_dataset)) # 483
# 15. 사전 훈련된 모델 내려받기
resnet18 = models.resnet18(pretrained=True) # 학습된 가중치를 사용
# 16. 사전 훈련된 모델의 파라미터 학습 유무 지정
def set_parameter_requires_grad(model, feature_extracting=True):
if feature_extracting:
for param in model.parameters():
param.requires_grad = False
set_parameter_requires_grad(resnet18)
# 17. ResNet18에 완전연결층 추가
resnet18.fc = nn.Linear(512,2) # 2는 클래스가 두 개라는 의미
# 18. 모델의 파라미터 값 확인
for name, param in resnet18.named_parameters():
if param.requires_grad:
print(name, param.data)
# fc.weight tensor([[ 0.0275, -0.0363, -0.0258, ..., 0.0214, -0.0426, -0.0397],
# [-0.0293, -0.0090, -0.0294, ..., 0.0064, -0.0259, -0.0146]])
# fc.bias tensor([-0.0380, 0.0403])
# 19. 모델 객체 생성 및 손실 함수 정의
model = models.resnet18(pretrained=True)
for param in model.parameters():
param.requires_grad = False
model.fc = torch.nn.Linear(512, 2)
for param in model.fc.parameters():
param.requires_grad = True
optimizer = torch.optim.Adam(model.fc.parameters())
cost = torch.nn.CrossEntropyLoss()
# print(model)
# ResNet(
# (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
# (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# (relu): ReLU(inplace=True)
# (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
# (layer1): Sequential(
# (0): BasicBlock(
# (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
# (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# (relu): ReLU(inplace=True)
# (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
# (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# (1): BasicBlock(
# (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
# (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# (relu): ReLU(inplace=True)
# (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
# (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
# )
# )
# (layer2): Sequential(
# ...
# )
# (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))
# (fc): Linear(in_features=512, out_features=2, bias=True)
# )
# 20. 모델 학습을 위한 함수 생성
def train_model(model, dataloaders, criterion, optimizer, device, num_epochs=13, is_train=True):
since = time.time()
acc_history = []
loss_history = []
best_acc= 0.0
for epoch in range(num_epochs):
print('Epoch{}/{}'.format(epoch, num_epochs-1))
print('-----------')
running_loss = 0.0
running_corrects = 0
for inputs, labels in dataloaders:
inputs = inputs.to(device)
labels = labels.to(device)
model.to(device)
optimizer.zero_grad()
outputs = model(inputs)
loss = criterion(outputs, labels)
_, preds = torch.max(outputs, 1)
loss.backward()
optimizer.step()
running_loss += loss.item() * inputs.size(0)
running_corrects += torch.sum(preds == labels.data)
epoch_loss = running_loss / len(dataloaders.dataset)
epoch_acc = running_corrects.double() / len(dataloaders.dataset)
print('Losss: {: .4f} Acc: {: .4f}'.format(epoch_loss, epoch_acc))
if epoch_acc > best_acc:
best_acc = epoch_acc
acc_history.append(epoch_loss)
loss_history.append(epoch_loss)
torch.save(model.state_dict(), os.path.join(r'C:\Users\032\SKT_AI\catanddog', '{0:0=2d}.pth'.format(epoch)))
print()
time_elapsed = time.time() - since
print('Training complete in {:.0f}m {: 0f}s'.format(time_elapsed // 60, time_elapsed % 60))
print('Best Acc: {:4f}'.format(best_acc))
return acc_history, loss_history
# 21. 파라미터 학습 결과를 옵테마이저에 전달
params_to_update = []
for namr, param in resnet18.named_parameters():
if param.requires_grad == True:
params_to_update.append(param)
print("\t", name)
optimizer = optim.Adam(params_to_update)
# 22. 모델 학습
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
criterion = nn.CrossEntropyLoss()
train_acc_hist, train_loss_hist = train_model(resnet18, train_loader, criterion, optimizer, device)
# 23. 테스트 데이터 호출 및 전처리
test_path = r'C:\Users\032\SKT_AI\catanddog'
transform = transforms.Compose(
[
transforms.Resize(224),
transforms.CenterCrop(224),
transforms.ToTensor(),
]
)
test_dataset = torchvision.datasets.ImageFolder(
root = test_path,
transform=transform
)
test_loader = torch.utils.data.DataLoader(
test_dataset,
batch_size=32,
num_workers=1,
shuffle=True
)
print(len(test_dataset))
# 24. 테스트 데이터 평가 함수 생성
def eval_model(model, dataloaders, device):
since = time.time()
acc_history = []
best_acc = 0.0
# 모델 경로 설정
save_path = r'C:\Users\032\SKT_AI\catanddog'
saved_models = glob.glob(os.path.join(save_path, '*.pth'))
saved_models.sort()
print('saved_models:', saved_models)
if not saved_models:
print("No saved models found!")
return []
for model_path in saved_models:
print('Loading model', model_path)
model.load_state_dict(torch.load(model_path))
model.eval()
model.to(device)
running_corrects = 0
for inputs, labels in dataloaders:
inputs = inputs.to(device)
labels = labels.to(device)
with torch.no_grad():
outputs = model(inputs)
_, preds = torch.max(outputs.data, 1)
preds[preds >= 0.5] = 1
preds[preds < 0.5] = 0
# running_corrects += preds.eq(labels.cpu()).int().sum()
running_corrects += preds.eq(labels).sum()
epoch_acc = running_corrects.double() / len(dataloaders.dataset)
print('ACC: {:.4f}'.format(epoch_acc))
if epoch_acc > best_acc:
best_acc = epoch_acc
acc_history.append(epoch_acc.item())
print()
time_elapsed = time.time() - since
print('Validation complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
print('Best Acc: {:4f}'.format(best_acc))
return acc_history
# 25. 테스트 데이터를 평가 함수에 적용
val_acc_hist = eval_model(resnet18, test_loader, device)
# 26. 훈련과 테스트 데이터의 정확도를 그래프로 확인
plt.plot(train_acc_hist)
plt.plot(val_acc_hist)
plt.show()
# 27. 훈련 데이터의 오차에 대한 그래프 확인
plt.plot(train_loss_hist)
plt.show()
# 28. 예측 이미지 출력을 위한 전처리 함수
def im_convert(tensor):
image = tensor.clone().detach().numpy()
image = image.transpose(1,2,0)
image = image * (np.array((0.5, 0.5, 0.5)) + np.array((0.5, 0.5, 0.5)))
image = image.clip(0,1)
return image
# 29. 개와 고양이 예측 결과 출력
classes = {0:'cat', 1:'dog'}
dataiter = iter(test_loader)
images, labels = next(iter(test_loader))
output = model(images)
_, preds = torch.max(output, 1)
fig = plt.figure(figsize=(25,4))
for idx in np.arange(20):
ax = fig.add_subplot(2, 10, idx+1, xticks=[], yticks=[])
plt.imshow(im_convert(images[idx]))
ax.set_title(classes[labels[idx].item()])
ax.set_title("{}({})".format(str(classes[preds[idx].item()]), str(classes[labels[idx].item()])), color=("green" if preds[idx]==labels[idx] else "red"))
plt.show()
plt.subplots_adjust(bottom=0.2, top=0.6, hspace=0)
더보기
483
fc.weight tensor([[-0.0220, 0.0208, -0.0145, ..., -0.0412, 0.0179, 0.0284],
[ 0.0056, 0.0089, -0.0391, ..., -0.0281, 0.0141, -0.0195]])
fc.bias tensor([-0.0267, 0.0308])
c:\Users\032\anaconda3\envs\torch_book\lib\site-packages\torchvision\models\_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
warnings.warn(
c:\Users\032\anaconda3\envs\torch_book\lib\site-packages\torchvision\models\_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=ResNet18_Weights.IMAGENET1K_V1`. You can also use `weights=ResNet18_Weights.DEFAULT` to get the most up-to-date weights.
warnings.warn(msg)
fc.bias
fc.bias
Epoch0/12
-----------
Losss: 0.5982 Acc: 0.7702
Epoch1/12
-----------
Losss: 0.5120 Acc: 0.7909
Epoch2/12
-----------
Losss: 0.5101 Acc: 0.7930
Epoch3/12
-----------
Losss: 0.5000 Acc: 0.7992
Epoch4/12
-----------
Losss: 0.4983 Acc: 0.8033
Epoch5/12
-----------
Losss: 0.4986 Acc: 0.7930
...
Best Acc: 0.815735
483
saved_models: ['C:\\Users\\032\\SKT_AI\\catanddog\\00.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\01.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\02.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\03.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\04.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\05.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\06.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\07.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\08.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\09.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\10.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\11.pth', 'C:\\Users\\032\\SKT_AI\\catanddog\\12.pth']
Loading model C:\Users\032\SKT_AI\catanddog\00.pth
Output is truncated. View as a scrollable element or open in a text editor. Adjust cell output settings...
C:\Users\032\AppData\Local\Temp\ipykernel_14920\2951342143.py:210: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
model.load_state_dict(torch.load(model_path))
ACC: 0.0559
Validation complete in 0m 4s
Best Acc: 0.055901'AI 및 Data Analysis > Code' 카테고리의 다른 글
| [OpenCV] 인터페이스 (디지털 영상 처리) (3) | 2025.01.21 |
|---|---|
| [YOLO] 시작 및 실행 (0) | 2025.01.18 |
| [scRNA] Preprocessing and Analysis (0) | 2024.12.14 |
| [MAST], [limma], [DESeq2] Common DEGs Analysis (0) | 2024.09.19 |
| [Seurat] [DESeq2] DEGs Analysis (0) | 2024.09.19 |
