USASI
SoftwareFramework

Liger Kernel

Project record

Maintained by LinkedIn137

Liger Kernel is an open-source collection of Triton GPU kernels for training large language models, developed at LinkedIn. It provides Hugging Face-compatible implementations of layers such as RMSNorm, RoPE, SwiGLU, and cross-entropy (including a fused linear cross-entropy), plus memory-efficient losses for post-training methods such as DPO, ORPO, and KTO.27

Last reviewedEntry updated Documented release Aug 2024

Availability and license

Overall availability

Public

Documented as available to the general public. Access conditions and license terms may still apply.16

Availability is separate from permission: read the license before using or redistributing.

Public materials checklist

Items for a framework under USASI rubric v0.1. Unknown means unassessed or insufficient evidence.
Public materials checklist for Liger Kernel
ItemStatusNotes and evidence
Source codeIs the source code publicly readable?Public1
DocumentationIs user documentation published?PublicA documentation site covers getting started, examples, and high- and low-level APIs.4
InstallationAre installation instructions or packages publicly available?PublicInstallable from PyPI (pip install liger-kernel), as a nightly package, or from source.26
Supported platformsAre supported operating systems or hardware documented?PublicThe README lists dependencies for NVIDIA CUDA (torch 2.1.2 or later, Triton 2.3.1 or later), AMD ROCm (torch 2.5.0 or later, Triton 3.0.0 or later), and Ascend NPU, plus optional CUDA-only cuTile and CuTe DSL backends (the latter targeting Hopper and Blackwell GPUs). It notes that the kernels inherit Triton's hardware compatibility.2
Release statusAre versioned releases published?PublicVersioned releases are published on GitHub and PyPI; v0.8.3 was released on September 16, 2026.56

What it is useful for

Reducing GPU memory use and increasing throughput when training or fine-tuning language models, either by patching a Hugging Face Transformers model with one line of code or by composing models from its modules. It works with PyTorch FSDP, DeepSpeed, and DDP, and is integrated in trainers such as Axolotl, LLaMA-Factory, TRL's SFTTrainer, and the Hugging Face Trainer.2

Organization context

U.S. eligibility

Project eligibility rests on documented governing or maintaining entities, not on contributors.

Eligible · basis: U.S.-governed project

The project is hosted in LinkedIn's GitHub organization, its LICENSE names LinkedIn Corporation as copyright holder, and LinkedIn Engineering describes it as LinkedIn's open-source project. LinkedIn Corporation is listed as a U.S. subsidiary in Exhibit 21 to Microsoft's Form 10-K for fiscal 2026 (see the LinkedIn record).1378

Assessed Sep 29, 2026

Sources

  1. 1.
    linkedin/Liger-Kernel (GitHub repository) (external site: github.com)

    LinkedIn · Repository · accessed Sep 29, 2026

  2. 2.
    Liger Kernel README (external site: raw.githubusercontent.com)

    LinkedIn · Repository · accessed Sep 29, 2026

  3. 3.
    Liger-Kernel LICENSE (external site: raw.githubusercontent.com)

    LinkedIn · License · accessed Sep 29, 2026

  4. 4.
    Liger-Kernel Docs (external site: linkedin.github.io)

    LinkedIn · Documentation · accessed Sep 29, 2026

  5. 5.
    Liger-Kernel releases (GitHub releases feed) (external site: github.com)

    LinkedIn · Release notes · accessed Sep 29, 2026

  6. 6.
    liger-kernel on PyPI (JSON metadata) (external site: pypi.org)

    Python Package Index · Other · accessed Sep 29, 2026

  7. 7.
    Liger-Kernel: Empowering an open source ecosystem of Triton Kernels for Efficient LLM Training (external site: linkedin.com)

    LinkedIn Engineering · Announcement · published Dec 5, 2024 · accessed Sep 29, 2026

  8. 8.
    Microsoft Corporation Form 10-K fiscal 2026, Exhibit 21 (Subsidiaries of Registrant) (external site: sec.gov)

    U.S. Securities and Exchange Commission (Microsoft filing) · Filing · published Jul 29, 2026 · accessed Sep 29, 2026

This listing is not an endorsement, a safety assessment, or a federal approval.

Support Us

Help keep USASI useful.

Optional. No USASI account required. Payment takes place on the linked provider’s website (Buy Me a Coffee).

About supporting this project