Skip to main content

OpenSSF Scorecard CodeQL

Intel® XPU Library for PyTorch* Ecosystem Projects

This project contains a set of plugins for PyTorch* ecosystem libraries which enable hardware acceleration on Intel® GPUs thru the xpu PyTorch* device backend. The goal of the project is to:

  • Facilitate enabling of the Intel® GPUs support across PyTorch* ecosystem projects
  • Provide the plugins till the support for Intel® GPUs will be accepted in the respective upstream projects

At the moment project provides plugins for the following frameworks:

Plugins

Intel® XPU plugin for TorchCodec

TorchCodec is a high-performance Python library designed for media processing (decoding and encoding) using PyTorch* tensors. Intel® XPU plugin for TorchCodec enables hardware acceleration for video operations (only decoding at the moment) on Linux. Both TorchCodec and Intel® plugin rely on the FFmpeg libraries for their operations which must be pre-installed on the system. Intel® plugin further assumes that FFmpeg is built with the VAAPI support.

TorchCodec will automatically load Intel® XPU plugin if it is installed on the system via Python packages entry point mechanism. After importing TorchCodec, pass XPU device to initialize TorchCodec decoder or encoder:

import torchcodec

decoder = torchcodec.decoders.VideoDecoder(
    "input.mp4", device="xpu:0")

Installation

Pre-built release wheels are available at PyPI. Installation requires PyTorch with enabled XPU support which can be fetched from https://download.pytorch.org/whl/xpu:

pip install torchlib-xpu \
  --extra-index-url https://download.pytorch.org/whl/xpu

Build from sources

  • Install uv

  • Install oneAPI 2026.1

  • Install FFmpeg development environment with enabled VAAPI hardware acceleration. For example:

    • By installing FFmpeg provided by your Linux distribution. For Ubuntu:
apt-get update && apt-get install -y \
    libavcodec-dev \
    libavdevice-dev \
    libavfilter-dev \
    libavformat-dev \
    libavutil-dev \
    libswresample-dev \
    libswscale-dev
  • By self-building FFmpeg from sources. The following example provides the minimal configuration required for hardware FFmpeg VAAPI codecs to be functional. For software fallback support, FFmpeg needs to be additionally built with enabled software codecs such as x264, x265, etc.:
git clone https://git.ffmpeg.org/ffmpeg.git && cd ffmpeg
./configure \
  --prefix=$HOME/_install \
  --libdir=$HOME/_install/lib \
  --disable-static \
  --disable-stripping \
  --disable-doc \
  --enable-shared \
  --enable-vaapi
make -j$(nproc) && make install

export PKG_CONFIG_PATH=$HOME/_install/lib/pkgconfig
export LD_LIBRARY_PATH=$HOME/_install/lib:$LD_LIBRARY_PATH
  • Build and install plugins supplied by Intel® XPU Library for PyTorch* Ecosystem Projects:
git clone https://github.com/intel/torchlib-xpu.git && cd torchlib-xpu

uv venv && uv pip install -e . \
  --index https://download.pytorch.org/whl/xpu -vv

Metadata

Release files for torchlib-xpu 0.5.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for torchlib-xpu 0.5.0
File Interpreter ABI Platform Actions
torchlib_xpu-0.5.0-py3-none-any.whl Python 3 none any Details for torchlib_xpu-0.5.0-py3-none-any.whl

Release files / torchlib_xpu-0.5.0-py3-none-any.whl

Download URL torchlib_xpu-0.5.0-py3-none-any.whl
Size 4.5 kB
Tags Python 3
SHA-256 checksum
How to use checksums
2107052878b6f4fcce8a6e4bdc2a570cc6a4a0f7975bfdcc17e0a1041585556f
BLAKE2b-256 checksum
How to use checksums
7aa0cb3202d404ba468bd3e6c888c872c86267a633be13d521ebb6dd66a1d378
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 15, 2026.

Transparency log

Security for torchlib-xpu 0.5.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

1 file in this release has been uploaded with provenance metadata.

Attestation for torchlib_xpu-0.5.0-py3-none-any.whl

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 15, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page