A Python library for GPU-accelerated tensor computation and dynamic neural network research, built on a tape-based autograd system. It enables developers and researchers to build, train, and customize deep learning models with imperative code execution.
Project overview
Provides a tape-based auto-differentiation system that allows arbitrary dynamic changes to network behavior during execution with zero lag.
Project type
Model Development · Infrastructure
Deployment
Refer to project documentation
License
License pending
Best for
Developers, researchers, and AI engineers who require a library for model-training-customization.
Key capabilities
Provides Tensors that can live either on the CPU or the GPU and accelerates scientific computation.
Uses reverse-mode auto-differentiation to allow changing the way your network behaves arbitrarily with zero lag or overhead.
A neural networks library deeply integrated with autograd designed for maximum flexibility.
Creates serializable and optimizable models from PyTorch code.
Provides a convenient extension API to write new neural network layers in C/C++ with minimal boilerplate.
Limitations and risks
Coding is required to use this library.
GPU acceleration is optional; tensors can live on the CPU or the GPU.
Building from source requires medium setup difficulty, C++ compilers, 30-60 minutes, and at least 10 GB of free disk space.
Getting started
Setup difficulty is medium. Users can install binaries via Conda or pip wheels, import torch, and build a neural network. Building from source requires C++ compilers, takes 30-60 minutes, and requires at least 10 GB of free disk space.
Alternatives and comparisons
Provides an autonomous research loop where an AI agent modifies training code, runs short experiments, evaluates results, and iterates automatically.
Curated lists of machine learning and deep learning frameworks, libraries, and data analysis tools categorized by programming language.
Provides a fast, dynamic language that combines ease of use with high performance for numerical and scientific computing.
GitHub project description: Tensors and Dynamic neural networks in Python with strong GPU acceleration
README: PyTorch is a Python package that provides two high-level features: - Tensor computation (like NumPy) with strong GPU acceleration - Deep neural networks built on a tape-based auto…
README: PyTorch is a library that consists of the following components:
README: we use a technique called reverse-mode auto-differentiation, which allows you to change the way your network behaves arbitrarily with zero lag or overhead.
README: PyTorch provides Tensors that can live either on the CPU or the GPU and accelerates the computation by a huge amount.