import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms
from torch.utils.data import DataLoader

class MLP(nn.Module):
    def __init__(self):
        super(MLP, self).__init__()
        self.layers = nn.Sequential(
            nn.Flatten(),
            nn.Linear(28*28, 128),
            nn.ReLU(),
            nn.Linear(128, 64),
            nn.ReLU(),
            nn.Linear(64, 10)
        )
    
    def forward(self, x):
        return self.layers(x)

transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (0.5,))])
train_data = datasets.MNIST(root='./data', train=True, download=True, transform=transform)
validation_data = datasets.MNIST(root='./data', train=False, download=True, transform=transform)
train_loader = DataLoader(train_data, batch_size=64, shuffle=True)
validation_loader = DataLoader(validation_data, batch_size=64, shuffle=False)

model = MLP()
optimizer = optim.SGD(model.parameters(), lr=0.01, momentum=0.9)
criterion = nn.CrossEntropyLoss()

def train(model, loader, optimizer, criterion):
    model.train()
    for data, target in loader:
        optimizer.zero_grad()
        output = model(data)
        loss = criterion(output, target)
        loss.backward()
        optimizer.step()

def validation(model, loader):
    model.eval()
    correct = 0
    total = 0
    with torch.no_grad():
        for data, target in loader:
            output = model(data)
            _, predicted = torch.max(output.data, 1)
            total += target.size(0)
            correct += (predicted == target).sum().item()
    print(f'Accuracy: {100 * correct / total}%')

for epoch in range(10):
    train(model, train_loader, optimizer, criterion)
    validation(model, validation_loader)

Accuracy: 93.2%
Accuracy: 94.93%
Accuracy: 95.71%
Accuracy: 96.67%
Accuracy: 96.29%
Accuracy: 97.23%
Accuracy: 97.18%
Accuracy: 97.45%
Accuracy: 97.56%
Accuracy: 97.62%

MLP w pytorch -- minimalny przykład¶