PyTorch – Open Source Machine Learning Framework

PyTorch is an open-source machine learning framework developed by Meta AI that has become the industry standard for deep learning research and production. Known for its dynamic computational graphs, intuitive Python interface, and excellent debugging capabilities, PyTorch powers many of today’s most advanced AI systems, from computer vision to natural language processing and generative AI.

Key Features

• Dynamic Computation Graphs – Define-by-run approach for flexible model building
• Pythonic Interface – Feels natural to Python developers
• GPU Acceleration – Seamless CUDA support for NVIDIA GPUs
• Distributed Training – Scale across multiple GPUs and machines
• TorchScript – Deploy models in production environments
• Rich Ecosystem – torchvision, torchaudio, torchtext libraries
• ONNX Export – Interoperability with other frameworks

Installation on Linux

# Install with pip (CPU only)
pip install torch torchvision torchaudio

# Install with CUDA 12.1 support
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

# Install with conda
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia

# Verify installation
python -c "import torch; print(torch.__version__); print(torch.cuda.is_available())"

Basic Usage Example

import torch
import torch.nn as nn
import torch.optim as optim

# Define a simple neural network
class SimpleNet(nn.Module):
    def __init__(self):
        super(SimpleNet, self).__init__()
        self.fc1 = nn.Linear(784, 128)
        self.fc2 = nn.Linear(128, 10)
    
    def forward(self, x):
        x = torch.relu(self.fc1(x))
        x = self.fc2(x)
        return x

# Create model and move to GPU if available
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = SimpleNet().to(device)

# Define loss and optimizer
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.001)

# Training loop
for epoch in range(10):
    optimizer.zero_grad()
    outputs = model(inputs)
    loss = criterion(outputs, labels)
    loss.backward()
    optimizer.step()

Key Libraries

• torchvision – Computer vision datasets, models, and transforms
• torchaudio – Audio processing and speech recognition
• torchtext – Natural language processing utilities
• PyTorch Lightning – High-level training framework
• Hugging Face Transformers – Pre-trained transformer models
• PyTorch Geometric – Graph neural networks

Use Cases

• Computer Vision – Image classification, object detection, segmentation
• Natural Language Processing – Text classification, translation, summarization
• Generative AI – GANs, diffusion models, language models
• Reinforcement Learning – Game playing, robotics, optimization
• Scientific Computing – Physics simulations, drug discovery

Download PyTorch