Autoencoder.py

import torch
print("PyTorch version:", torch.__version__)
print("CUDA available:", torch.cuda.is_available())
print("GPU name:", torch.cuda.get_device_name(0))
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms
from torch.utils.data import DataLoader
from datasets import load_dataset

batch_size = 32
transform = transforms.Compose([
    transforms.Resize((128, 128)),
    transforms.ToTensor()
])

ds = load_dataset("saitsharipov/CelebA-HQ")


class CelebAHQDataset(torch.utils.data.Dataset):
    def __init__(self, hf_dataset, transform=None):
        self.dataset = hf_dataset
        self.transform = transform

    def __len__(self):
        return len(self.dataset)

    def __getitem__(self, idx):
        img = self.dataset[idx]["image"]  # Hugging Face provides PIL images
        if self.transform:
            img = self.transform(img)
        return img, 0  # dummy label (autoencoder doesn’t need labels)

# Train/test split
train_data = CelebAHQDataset(ds["train"], transform=transform)

train_loader = DataLoader(train_data, batch_size=16, shuffle=True)

class Autoencoder(nn.Module):
    def __init__(self):
        super(Autoencoder, self).__init__()
        self.encoder = nn.Sequential(
            nn.Conv2d(3,32,3, stride=2, padding=1),
            nn.ReLU(),
            nn.Conv2d(32,64,3, stride=2, padding=1),
            nn.ReLU(),
            nn.Conv2d(64, 128, 3, stride=2, padding=1),
            nn.ReLU(),
            nn.Conv2d(128, 256, 3, stride=2, padding=1)
        )

        self.decoder = nn.Sequential(
            nn.ConvTranspose2d(256, 128, 3, stride=2, padding=1, output_padding=1), # 16x16
            nn.ReLU(),
            nn.ConvTranspose2d(128, 64, 3, stride=2, padding=1, output_padding=1),  # 32x32
            nn.ReLU(),
            nn.ConvTranspose2d(64, 32, 3, stride=2, padding=1, output_padding=1),   # 64x64
            nn.ReLU(),
            nn.ConvTranspose2d(32, 3, 3, stride=2, padding=1, output_padding=1),    # 128x128
            nn.Sigmoid()
        )

    def forward(self, x):
        x = self.encoder(x)
        x = self.decoder(x)
        return x
    
model = Autoencoder()
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)
criterion = nn.L1Loss()
optimizer = optim.Adam(model.parameters(), lr=1e-4)

num_epochs = 50
for epoch in range(num_epochs):
    model.train()
    running_loss = 0.0
    for data in train_loader:
        imgs, _ = data
        imgs = imgs.to(device)

        optimizer.zero_grad()
        outputs = model(imgs)
        loss = criterion(outputs, imgs)
        
        loss.backward()
        optimizer.step()
        running_loss += loss.item() * imgs.size(0)

    epoch_loss = running_loss / len(train_loader.dataset)
    print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item(): .4f}')

torch.save(model.state_dict(), "celeba_autoencoder.pth")
print("Model saved!")