Skip to content
FindOpenSource

PyTorch

Open source machine learning framework that accelerates the path from research prototyping to production deployment.

Overview

PyTorch is a Python library built around two core capabilities: a NumPy-like tensor library with strong GPU acceleration, and a deep learning framework built on a tape-based automatic differentiation system (autograd) that tracks operations on tensors to compute gradients. Its `torch.nn` module builds neural network layers on top of autograd, `torch.jit` (TorchScript) can compile PyTorch code into serializable, optimizable models, and existing Python libraries like NumPy, SciPy, and Cython can be used to extend it.

PyTorch is typically used either as a GPU-accelerated replacement for NumPy, or as a deep learning research platform, and its design favors an imperative, Python-first style - you write and debug models as regular Python code rather than a separate static graph definition.

It's a fit for researchers and engineers building and training neural networks who want a flexible, Python-native tensor and autograd library rather than a higher-level, more opinionated deep learning framework.

Categories
AI & Machine Learning
Keywords
machine-learningdeep-learningneural-networksai
Languages
Python, C++
License
BSD-3-Clause

Spotted an error? Suggest an edit on GitHub.