pytorch基础学习

pytorch基础学习四

 

In [2]:

import torch
from torch import nn
from torch.utils.data import DataLoader
from torchvision import datasets
from torchvision.transforms import ToTensor

 

In [3]:

# 处理数据
# 从开放数据集下载训练数据。
training_data = datasets.FashionMNIST(
    root="data",
    train=True,
    download=True,
    transform=ToTensor(),
)

 

In [4]:

# 从开放数据集下载测试数据。
test_data = datasets.FashionMNIST(
    root="data",
    train=False,
    download=True,
    transform=ToTensor(),
)
batch_size = 64

 

In [5]:

# 创建数据加载器。
train_dataloader = DataLoader(training_data, batch_size=batch_size)
test_dataloader = DataLoader(test_data, batch_size=batch_size)

 

In [6]:

for X, y in test_dataloader:
    print(f"Shape of X [N, C, H, W]: {X.shape}")
    print(f"Shape of y: {y.shape}{y.dtype}")
    break

 

 

Shape of X [N, C, H, W]: torch.Size([64, 1, 28, 28])
Shape of y: torch.Size([64]) torch.int64

 

In [7]:

# 创建模型
# 获取cpu或gpu设备进行训练
device = "cuda" if torch.cuda.is_available() else "cpu"
print(f"Using {device} device")

 

 

Using cpu device

 

In [8]:

# 定义模型
class NeuralNetwork(nn.Module):
    def __init__(self):
        super(NeuralNetwork, self).__init__()
        self.flatten = nn.Flatten()
        self.linear_relu_stack = nn.Sequential(
            nn.Linear(28*28, 512),
            nn.ReLU(),
            nn.Linear(512, 512),
            nn.ReLU(),
            nn.Linear(512, 10)
        )

    def forward(self, x):
        x = self.flatten(x)
        logits = self.linear_relu_stack(x)
        return logits

model = NeuralNetwork().to(device)
print(model)

 

 

NeuralNetwork(
  (flatten): Flatten(start_dim=1, end_dim=-1)
  (linear_relu_stack): Sequential(
    (0): Linear(in_features=784, out_features=512, bias=True)
    (1): ReLU()
    (2): Linear(in_features=512, out_features=512, bias=True)
    (3): ReLU()
    (4): Linear(in_features=512, out_features=10, bias=True)
  )
)

 

In [9]:

# 优化模型参数
loss_fn = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr=1e-3)

 

In [10]:

# 在单个训练循环中，模型对训练数据集进行预测（分批输入），并反向传播预测误差以调整模型的参数。
def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)
    model.train()
    for batch, (X, y) in enumerate(dataloader):
        X, y = X.to(device), y.to(device)

        # 计算预测误差
        pred = model(X)
        loss = loss_fn(pred, y)

        # 反向传播
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        if batch % 100 == 0:
            loss, current = loss.item(), batch * len(X)
            print(f"loss: {loss:>7f}  [{current:>5d}/{size:>5d}]")

 

In [11]:

# 根据测试数据集检查模型的性能，以确保它正在学习。
def test(dataloader, model, loss_fn):
    size = len(dataloader.dataset)
    num_batches = len(dataloader)
    model.eval()
    test_loss, correct = 0, 0
    with torch.no_grad():
        for X, y in dataloader:
            X, y = X.to(device), y.to(device)
            pred = model(X)
            test_loss += loss_fn(pred, y).item()
            correct += (pred.argmax(1) == y).type(torch.float).sum().item()
    test_loss /= num_batches
    correct /= size
    print(f"Test Error: \n Accuracy: {(100*correct):>0.1f}%, Avg loss: {test_loss:>8f}\n")

 

In [12]:

# 训练过程在多次迭代（epochs）中进行。在每个时期，模型都会学习参数以做出更好的预测。我们在每个时期打印模型的准确性和损失；
# 我们希望看到每个 epoch 的准确率增加和损失减少。
epochs = 5
for t in range(epochs):
    print(f"Epoch {t+1}\n-------------------------------")
    train(train_dataloader, model, loss_fn, optimizer)
    test(test_dataloader, model, loss_fn)
print("Done!")

 

 

Epoch 1
-------------------------------
loss: 2.307462  [    0/60000]
loss: 2.289459  [ 6400/60000]
loss: 2.273360  [12800/60000]
loss: 2.263474  [19200/60000]
loss: 2.252384  [25600/60000]
loss: 2.221936  [32000/60000]
loss: 2.228031  [38400/60000]
loss: 2.192849  [44800/60000]
loss: 2.194643  [51200/60000]
loss: 2.164009  [57600/60000]
Test Error: 
 Accuracy: 48.0%, Avg loss: 2.155143 

Epoch 2
-------------------------------
loss: 2.170278  [    0/60000]
loss: 2.157974  [ 6400/60000]
loss: 2.101683  [12800/60000]
loss: 2.111527  [19200/60000]
loss: 2.070840  [25600/60000]
loss: 2.006484  [32000/60000]
loss: 2.034235  [38400/60000]
loss: 1.955799  [44800/60000]
loss: 1.969466  [51200/60000]
loss: 1.894076  [57600/60000]
Test Error: 
 Accuracy: 61.5%, Avg loss: 1.887733 

Epoch 3
-------------------------------
loss: 1.930098  [    0/60000]
loss: 1.895441  [ 6400/60000]
loss: 1.777315  [12800/60000]
loss: 1.807420  [19200/60000]
loss: 1.716668  [25600/60000]
loss: 1.657058  [32000/60000]
loss: 1.677540  [38400/60000]
loss: 1.579168  [44800/60000]
loss: 1.611868  [51200/60000]
loss: 1.495966  [57600/60000]
Test Error: 
 Accuracy: 62.3%, Avg loss: 1.512710 

Epoch 4
-------------------------------
loss: 1.595690  [    0/60000]
loss: 1.550450  [ 6400/60000]
loss: 1.395835  [12800/60000]
loss: 1.455848  [19200/60000]
loss: 1.352183  [25600/60000]
loss: 1.337337  [32000/60000]
loss: 1.349648  [38400/60000]
loss: 1.279088  [44800/60000]
loss: 1.323315  [51200/60000]
loss: 1.211475  [57600/60000]
Test Error: 
 Accuracy: 63.5%, Avg loss: 1.239858 

Epoch 5
-------------------------------
loss: 1.334896  [    0/60000]
loss: 1.308266  [ 6400/60000]
loss: 1.136204  [12800/60000]
loss: 1.234191  [19200/60000]
loss: 1.116831  [25600/60000]
loss: 1.132986  [32000/60000]
loss: 1.156611  [38400/60000]
loss: 1.099791  [44800/60000]
loss: 1.145944  [51200/60000]
loss: 1.056027  [57600/60000]
Test Error: 
 Accuracy: 64.5%, Avg loss: 1.077008 

Done!

 

In [13]:

# 保存模型
# 保存模型的常用方法是序列化内部状态字典（包含模型参数）。
torch.save(model.state_dict(), "model.pth")
print("Saved PyTorch Model State to model.pth")

 

 

Saved PyTorch Model State to model.pth

 

In [14]:

# 加载模型
# 加载模型的过程包括重新创建模型结构并将状态字典加载到其中。
model = NeuralNetwork()
model.load_state_dict(torch.load("model.pth"))

 

Out[14]:

<All keys matched successfully>

 

In [15]:

# 利用该模型进行预测。
classes = [
    "T-shirt/top",
    "Trouser",
    "Pullover",
    "Dress",
    "Coat",
    "Sandal",
    "Shirt",
    "Sneaker",
    "Bag",
    "Ankle boot",
]
model.eval()
x, y = test_data[0][0], test_data[0][1]
with torch.no_grad():
    pred = model(x)
    predicted, actual = classes[pred[0].argmax(0)], classes[y]
    print(f'Predicted: "{predicted}", Actual: "{actual}"')

 

 

Predicted: "Ankle boot", Actual: "Ankle boot"