Skip to content
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Branch: master
Clone or download
SsnL and facebook-github-bot Update multiprocessing note now that shared CUDA tensors are refcount…
…ed (#19904)

The mp notes are not updated after #16854. (The torch.multiprocessing page is.)
Pull Request resolved: #19904

Differential Revision: D15509661

Pulled By: soumith

fbshipit-source-id: 7c11e14a6c804498dda3adbf19710e63e6a564a0
Latest commit 83fe928 May 26, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
.circleci Fix pytorch_macos_10_13_py3_test (#20944) May 25, 2019
.ctags.d Add a .ctags.d/ toplevel directory (#18827) Apr 4, 2019
.github Update issue templates Sep 26, 2018
.jenkins Fix pytorch_macos_10_13_py3_test (#20944) May 25, 2019
aten Fix advanced indexing on "huge" Tensors (#20919) May 24, 2019
benchmarks Benchmark repeat op. (#20016) May 20, 2019
binaries Lightweight at-most-once logging for API usage (#20745) May 24, 2019
c10 Use Device instead of Backend in TensorIterator (#20690) May 24, 2019
caffe2 fix bug of not using get_score_cls_index in BoxWithNMSLimitOp (#20868) May 25, 2019
cmake Don't split 256-bit AVX2 load/store intrinsics (#20609) May 17, 2019
docker Add llvm8 installation step. (#20879) May 24, 2019
docs Update multiprocessing note now that shared CUDA tensors are refcount… May 26, 2019
modules Optimize SoftmaxOp on CPU (#18635) Apr 11, 2019
scripts Move torchvision install out of onnx test script May 24, 2019
submodules 'Re-sync with internal repository' (#12652) Oct 15, 2018
test allow pass ordered dict for nn sequential (#20796) May 25, 2019
third_party Updating submodules May 24, 2019
tools Resend "Split ATen/Parallel into interface and backend" (#20825) May 24, 2019
torch allow pass ordered dict for nn sequential (#20796) May 25, 2019
.clang-format Updates to .clang-format (#7683) May 18, 2018
.clang-tidy Enable all clang-tidy performance checks (#15198) Dec 14, 2018
.dockerignore Add .dockerignore. (#3333) Oct 28, 2017
.flake8 Remove flake8 E303 (too many blank lines) (#20225) May 7, 2019
.gitattributes add .gitattributes for EOL conversion. (#9813) Aug 1, 2018
.gitignore Reduce set of build/tests which run on PRs. (#20930) May 24, 2019
.gitmodules Use ignore=dirty in submodules. (#20135) May 7, 2019
.travis.aten.yml use flake8-mypy (#17721) Mar 7, 2019
.travis.yml Run shellcheck on Jenkins scripts (#18874) Apr 15, 2019
CITATION Improve the recommended citation May 21, 2019
CMakeLists.txt Leave it as an option for whether to colorize output during build (#2… May 24, 2019
CODEOWNERS Update CODEOWNERS (#17720) Mar 6, 2019 CUDA_NVCC_EXECUTABLE is not needed, as nvcc is in PATH (#18958) Apr 5, 2019
LICENSE Move copyright lines back to NOTICE file, fixes #6911 (#8310) Jun 12, 2018
Makefile Fix python support problems caused by building script errors. Apr 25, 2017
NOTICE Move copyright lines back to NOTICE file, fixes #6911 (#8310) Jun 12, 2018 tensor_illustration with correct numbers and better fonts for README … May 24, 2019 Change output_declarations in to be a NamedTuple (#… Feb 23, 2018
mypy-files.txt Change output_declarations in to be a NamedTuple (#… Feb 23, 2018
mypy.ini Change output_declarations in to be a NamedTuple (#… Feb 23, 2018
requirements.txt Remove protobuf require and use requirements.txt (#10771) Aug 24, 2018 Update version strings to 1.2 May 24, 2019
ubsan.supp restore hidden visibility by default for Linux builds (#20461) May 20, 2019

PyTorch Logo

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 autograd system

You can reuse your favorite Python packages such as NumPy, SciPy and Cython to extend PyTorch when needed.

System 2.7 3.5 3.6
Linux CPU Build Status Build Status
Linux GPU Build Status Build Status
Windows CPU / GPU Build Status
Linux (ppc64le) CPU Build Status Build Status
Linux (ppc64le) GPU Build Status Build Status

See also the HUD.

More About PyTorch

At a granular level, PyTorch is a library that consists of the following components:

Component Description
torch a Tensor library like NumPy, with strong GPU support
torch.autograd a tape-based automatic differentiation library that supports all differentiable Tensor operations in torch
torch.jit a compilation stack (TorchScript) to create serializable and optimizable models from PyTorch code
torch.nn a neural networks library deeply integrated with autograd designed for maximum flexibility
torch.multiprocessing Python multiprocessing, but with magical memory sharing of torch Tensors across processes. Useful for data loading and Hogwild training
torch.utils DataLoader and other utility functions for convenience

Usually one uses PyTorch either as:

  • a replacement for NumPy to use the power of GPUs.
  • a deep learning research platform that provides maximum flexibility and speed.

Elaborating further:

A GPU-Ready Tensor Library

If you use NumPy, then you have used Tensors (a.k.a ndarray).

Tensor illustration

PyTorch provides Tensors that can live either on the CPU or the GPU, and accelerates the computation by a huge amount.

We provide a wide variety of tensor routines to accelerate and fit your scientific computation needs such as slicing, indexing, math operations, linear algebra, reductions. And they are fast!

Dynamic Neural Networks: Tape-Based Autograd

PyTorch has a unique way of building neural networks: using and replaying a tape recorder.

Most frameworks such as TensorFlow, Theano, Caffe and CNTK have a static view of the world. One has to build a neural network, and reuse the same structure again and again. Changing the way the network behaves means that one has to start from scratch.

With PyTorch, 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. Our inspiration comes from several research papers on this topic, as well as current and past work such as torch-autograd, autograd, Chainer, etc.

While this technique is not unique to PyTorch, it's one of the fastest implementations of it to date. You get the best of speed and flexibility for your crazy research.

Dynamic graph

Python First

PyTorch is not a Python binding into a monolithic C++ framework. It is built to be deeply integrated into Python. You can use it naturally like you would use NumPy / SciPy / scikit-learn etc. You can write your new neural network layers in Python itself, using your favorite libraries and use packages such as Cython and Numba. Our goal is to not reinvent the wheel where appropriate.

Imperative Experiences

PyTorch is designed to be intuitive, linear in thought and easy to use. When you execute a line of code, it gets executed. There isn't an asynchronous view of the world. When you drop into a debugger, or receive error messages and stack traces, understanding them is straightforward. The stack trace points to exactly where your code was defined. We hope you never spend hours debugging your code because of bad stack traces or asynchronous and opaque execution engines.

Fast and Lean

PyTorch has minimal framework overhead. We integrate acceleration libraries such as Intel MKL and NVIDIA (cuDNN, NCCL) to maximize speed. At the core, its CPU and GPU Tensor and neural network backends (TH, THC, THNN, THCUNN) are mature and have been tested for years.

Hence, PyTorch is quite fast – whether you run small or large neural networks.

The memory usage in PyTorch is extremely efficient compared to Torch or some of the alternatives. We've written custom memory allocators for the GPU to make sure that your deep learning models are maximally memory efficient. This enables you to train bigger deep learning models than before.

Extensions Without Pain

Writing new neural network modules, or interfacing with PyTorch's Tensor API was designed to be straightforward and with minimal abstractions.

You can write new neural network layers in Python using the torch API or your favorite NumPy-based libraries such as SciPy.

If you want to write your layers in C/C++, we provide a convenient extension API that is efficient and with minimal boilerplate. There is no wrapper code that needs to be written. You can see a tutorial here and an example here.



Commands to install from binaries via Conda or pip wheels are on our website:

NVIDIA Jetson platforms

Python wheels for NVIDIA's Jetson Nano, Jetson TX2, and Jetson AGX Xavier are available via the following URLs:

They requires JetPack 4.2 and above and are maintained by @dusty-nv

From Source

If you are installing from source, we highly recommend installing an Anaconda environment. You will get a high-quality BLAS library (MKL) and you get controlled dependency versions regardless of your Linux distro.

Once you have Anaconda installed, here are the instructions.

If you want to compile with CUDA support, install

If you want to disable CUDA support, export environment variable NO_CUDA=1. Other potentially useful environment variables may be found in

If you are building for NVIDIA's Jetson platforms (Jetson Nano, TX1, TX2, AGX Xavier), Instructions to are available here

Install Dependencies


conda install numpy ninja pyyaml mkl mkl-include setuptools cmake cffi typing

On Linux

# Add LAPACK support for the GPU if needed
conda install -c pytorch magma-cuda90 # or [magma-cuda92 | magma-cuda100 ] depending on your cuda version

Get the PyTorch Source

git clone --recursive
cd pytorch
# if you are updating an existing checkout
git submodule sync
git submodule update --init --recursive

Install PyTorch

On Linux

export CMAKE_PREFIX_PATH=${CONDA_PREFIX:-"$(dirname $(which conda))/../"}
python install

On macOS

export CMAKE_PREFIX_PATH=${CONDA_PREFIX:-"$(dirname $(which conda))/../"}
MACOSX_DEPLOYMENT_TARGET=10.9 CC=clang CXX=clang++ python install

On Windows

At least Visual Studio 2017 Update 3 (version 15.3.3 with the toolset 14.11) and NVTX are needed.

If the version of Visual Studio 2017 is higher than 15.4.5, installing of "VC++ 2017 version 15.4 v14.11 toolset" is strongly recommended.
If the version of Visual Studio 2017 is lesser than 15.3.3, please update Visual Studio 2017 to the latest version along with installing "VC++ 2017 version 15.4 v14.11 toolset".
There is no guarantee of the correct building with VC++ 2017 toolsets, others than version 15.4 v14.11.
"VC++ 2017 version 15.4 v14.11 toolset" might be installed onto already installed Visual Studio 2017 by running its installation once again and checking the corresponding checkbox under "Individual components"/"Compilers, build tools, and runtimes".

NVTX is a part of CUDA distributive, where it is called "Nsight Compute". For installing it onto already installed CUDA run CUDA installation once again and check the corresponding checkbox. Be sure that CUDA with Nsight Compute is installed after Visual Studio 2017.

REM [Optional] The following two lines are needed for Python 2.7, but the support for it is very experimental.
set MSSdk=1

set CMAKE_GENERATOR=Visual Studio 15 2017 Win64

REM Run "Visual Studio 2017 Developer Command Prompt"
for /f "usebackq tokens=*" %i in (`"%ProgramFiles(x86)%\Microsoft Visual Studio\Installer\vswhere.exe" -version [15^,16^) -products * -latest -property installationPath`) do call "%i\VC\Auxiliary\Build\vcvarsall.bat" x64 -vcvars_ver=14.11

python install

Docker Image

Dockerfile is supplied to build images with cuda support and cudnn v7. You can pass -e PYTHON_VERSION=x.y flag to specify which Python version is to be used by Miniconda, or leave it unset to use the default. Build from pytorch repo directory as docker needs to copy git repo into docker filesystem while building the image.

docker build -t pytorch -f docker/pytorch/Dockerfile .

You can also pull a pre-built docker image from Docker Hub and run with nvidia-docker, but this is not currently maintained and will pull PyTorch 0.2.

nvidia-docker run --rm -ti --ipc=host pytorch/pytorch:latest

Please note that PyTorch uses shared memory to share data between processes, so if torch multiprocessing is used (e.g. for multithreaded data loaders) the default shared memory segment size that container runs with is not enough, and you should increase shared memory size either with --ipc=host or --shm-size command line options to nvidia-docker run.

Building the Documentation

To build documentation in various formats, you will need Sphinx and the readthedocs theme.

cd docs/
pip install -r requirements.txt

You can then build the documentation by running make <format> from the docs/ folder. Run make to get a list of all available output formats.

Previous Versions

Installation instructions and binaries for previous PyTorch versions may be found on our website.

Getting Started

Three pointers to get you started:


  • forums: discuss implementations, research, etc.
  • GitHub issues: bug reports, feature requests, install issues, RFCs, thoughts, etc.
  • Slack: The PyTorch Slack hosts a primary audience of moderate to experienced PyTorch users and developers for general chat, online discussions, collaboration etc. If you are a beginner looking for help, the primary medium is PyTorch Forums. If you need a slack invite, please fill this form:
  • newsletter: no-noise, one-way email newsletter with important announcements about pytorch. You can sign-up here:

Releases and Contributing

PyTorch has a 90 day release cycle (major releases). Please let us know if you encounter a bug by filing an issue.

We appreciate all contributions. If you are planning to contribute back bug-fixes, please do so without any further discussion.

If you plan to contribute new features, utility functions or extensions to the core, please first open an issue and discuss the feature with us. Sending a PR without discussion might end up resulting in a rejected PR, because we might be taking the core in a different direction than you might be aware of.

The Team

PyTorch is a community driven project with several skillful engineers and researchers contributing to it.

PyTorch is currently maintained by Adam Paszke, Sam Gross, Soumith Chintala and Gregory Chanan with major contributions coming from hundreds of talented individuals in various forms and means. A non-exhaustive but growing list needs to mention: Trevor Killeen, Sasank Chilamkurthy, Sergey Zagoruyko, Adam Lerer, Francisco Massa, Alykhan Tejani, Luca Antiga, Alban Desmaison, Andreas Kopf, James Bradbury, Zeming Lin, Yuandong Tian, Guillaume Lample, Marat Dukhan, Natalia Gimelshein, Christian Sarofeen, Martin Raison, Edward Yang, Zachary Devito.

Note: this project is unrelated to hughperkins/pytorch with the same name. Hugh is a valuable contributor in the Torch community and has helped with many things Torch and PyTorch.


PyTorch is BSD-style licensed, as found in the LICENSE file.

You can’t perform that action at this time.