import torch
import torchvision
import torchvision.transforms as transforms
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from safetensors import safe_open
from safetensors.torch import save_file

# 데이터셋 불러오기
transform = transforms.Compose(
    [transforms.ToTensor(),
     transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])

batch_size = 4

trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
                                        download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=batch_size,
                                          shuffle=True, num_workers=2)

testset = torchvision.datasets.CIFAR10(root='./data', train=False,
                                       download=True, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=batch_size,
                                         shuffle=False, num_workers=2)

classes = ('plane', 'car', 'bird', 'cat',
           'deer', 'dog', 'frog', 'horse', 'ship', 'truck')

# 합성곱 신경망 만들기
class Net(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(3, 6, 5)
        self.pool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(6, 16, 5)
        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(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = torch.flatten(x, 1) # 배치를 제외한 모든 차원을 평탄화(flatten)
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        x = self.fc3(x)
        return x


net = Net()

# 손실 함수와 오티마이져
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)

# 학습하기
for epoch in range(2):   # 데이터셋을 수차례 반복합니다.

    running_loss = 0.0
    for i, data in enumerate(trainloader, 0):
        # [inputs, labels]의 목록인 data로부터 입력을 받은 후;
        inputs, labels = data

        # 변화도(Gradient) 매개변수를 0으로 만들고
        optimizer.zero_grad()

        # 순전파 + 역전파 + 최적화를 한 후
        outputs = net(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()

        # 통계를 출력합니다.
        running_loss += loss.item()
        if i % 2000 == 1999:    # print every 2000 mini-batches
            print(f'[{epoch + 1}, {i + 1:5d}] loss: {running_loss / 2000:.3f}')
            running_loss = 0.0

print('Finished Training')

# 모델 저장하기 
PATH = './cifar_net.pth'
torch.save(net.state_dict(), PATH) # Not safe way

save_file(net.state_dict(), "model.safetensors")

# 모델 불러오기
tensors = {}
with safe_open("model.safetensors", framework="pt", device="cpu") as f:
   for key in f.keys():
       tensors[key] = f.get_tensor(key)