{"data":{"slug":"linkedin-liger-kernel","name":"Liger-Kernel","tagline":"Efficient Triton Kernels for LLM Training","github_url":"https://github.com/linkedin/Liger-Kernel","owner":"linkedin","repo":"Liger-Kernel","owner_avatar_url":"https://avatars.githubusercontent.com/u/357098?v=4","primary_language":"Python","stars":6555,"forks":573,"topics":["finetuning","gemma2","hacktoberfest","llama","llama3","llm-training","llms","mistral","phi3","triton","triton-kernels"],"archived":false,"github_pushed_at":"2026-08-07T08:48:09+00:00","maintenance_label":"Very active","url":"https://www.graphcanon.com/tools/linkedin-liger-kernel","markdown_url":"https://www.graphcanon.com/tools/linkedin-liger-kernel.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/linkedin-liger-kernel","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=linkedin-liger-kernel","description":"Efficient Triton Kernels for LLM Training","homepage_url":"https://linkedin.github.io/Liger-Kernel/","license":"BSD-2-Clause","open_issues":190,"watchers":55,"ai_summary":"Provides optimized Triton kernels designed to speed up the training process of large language models (LLMs). Supports ROCm PyTorch installation and offers customization options through installation flags.","readme_excerpt":"# On ROCm, install ROCm PyTorch first from the PyTorch ROCm index.\npip install -e .\n\n---\n\n# Or install cuTile with the optional tileiras compiler\npip install -e \".[cutile-tileiras]\"\n\n---\n\n## Getting Started\n\nThere are a couple of ways to apply Liger kernels, depending on the level of customization required.\n\n---\n\n## Contributing, Acknowledgements, and License\n\n- [Contributing Guidelines](https://github.com/linkedin/Liger-Kernel/blob/main/docs/contributing.md)\n- [Acknowledgements](https://github.com/linkedin/Liger-Kernel/blob/main/docs/acknowledgement.md)\n- [License Information](https://github.com/linkedin/Liger-Kernel/blob/main/docs/license.md)","github_created_at":"2024-08-06T17:47:52+00:00","created_at":"2026-07-11T10:37:57.417956+00:00","updated_at":"2026-08-07T12:00:53.578518+00:00","categories":[{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"}],"tags":[{"slug":"finetuning","name":"finetuning"},{"slug":"gemma2","name":"gemma2"},{"slug":"llama","name":"llama"},{"slug":"mistral","name":"mistral"},{"slug":"phi3","name":"phi3"},{"slug":"triton-kernels","name":"triton-kernels"}],"trust":{"provenance":{"is_fork":false,"github_id":838970603,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-08-07T12:00:52.779Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":1,"days_since_push":0,"last_release_at":"2026-07-23T00:01:09Z"},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T10:37:58.820Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-08-07T12:00:53.254Z"},"languages":{"value":["python"],"source":"github.language+pyproject.toml","observed_at":"2026-08-07T12:00:53.254Z"},"license_spdx":{"value":"BSD-2-Clause","source":"github.license","observed_at":"2026-08-07T12:00:53.254Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When enhancing training speed of large language models with ROCm-compatible hardware.","For projects requiring customization through install flags, like enabling cuTile with tileiras."],"when_not_to_use":["Avoid if only CUDA environments are supported, as Liger-Kernel emphasizes ROCm compatibility.","Skip for simple setup requirements; prefer more streamlined tools without extensive customization options."],"source":"enrich:decision_facts","observed_at":"2026-07-12T13:35:19.873Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"Optimized Triton kernels for accelerating LLM training, especially on ROCm PyTorch installations."}]}}