AI Potluck
Model components / Fine-tuning code

LLaMA-Factory

Unknown

LLaMA-Factory is a toolkit for fine-tuning 100+ open LLM and vision-language architectures, including LLaMA, Qwen, DeepSeek, Gemma, and GLM. It covers continued pre-training, supervised fine-tuning, reward modeling, and preference or RL methods such as PPO, DPO, KTO, and ORPO, across full, LoRA, and QLoRA tuning. A CLI and the Gradio-based LLaMA Board web UI run the same pipeline without custom code, and the project appeared as an ACL 2024 paper.

Verified 2026-08-09 via GitHub and the LICENSE body.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI)
source
public(hiyouga/LLaMA-Factory)
core-gated
ungated

Apache-2.0, an OSI license, with the full source public including the web UI, and no features held back for a paid tier.

  • https://github.com/hiyouga/LLaMA-Factory recorded 2026-08-09

    Redirects (301) to the renamed repository https://github.com/hiyouga/LlamaFactory (same owner, repo id 646410686, created 2023-05-28). The rendered README states "This repository is licensed under the Apache-2.0 License." The top-level tree is .ai, .claude, .github, assets, data, docker, docs, examples, requirements, scripts, src, tests, tests_v1 plus root files including LICENSE — no enterprise, commercial, or paid-tier directory, and src/ ships the whole toolkit including the CLI and the Gradio-based LLaMA Board web UI. Not archived; sidebar reports 73,920 stars and 9,042 forks.

  • https://raw.githubusercontent.com/hiyouga/LlamaFactory/main/LICENSE recorded 2026-08-09

    The full standard Apache License, Version 2.0 text — definitions through the appendix boilerplate — with no additional restriction, exception, or commons-clause layered on top.

  • https://pypi.org/project/llamafactory/ recorded 2026-08-09

    Project page for "llamafactory 0.9.5" lists License expression Apache-2.0 and points Homepage/Repository at https://github.com/hiyouga/LLaMA-Factory, confirming the published package matches the GitHub source.

Adoption

2 medium confidence
2.0

PyPI downloads of the llamafactory package spiked and then returned to their earlier level rather than collapsing: 14,311 in February, 27,761 in March and 30,043 in April, then 233,960 in May and 779,461 in June, falling back to 100,640 in July and 21,867 over the trailing 30 days. Steady state on either side of the spike is 20-30K a month, which puts reach in the 10K-100K band. Stars and forks kept growing throughout, and the project stands at 73,939 stars, so this is a healthy project with modest package installs rather than a fading one. One undercount is known but cannot be quantified: the project is commonly run from a git clone and its web UI rather than installed from PyPI, and GitHub releases capture almost none of that either, at 5,242 asset downloads across 30 releases. No available signal measures the clone channel.

Capability

4 medium confidence
4.0

Breadth rather than scale. The project describes itself as unified efficient fine-tuning of more than 100 LLMs and VLMs, and covers LoRA, QLoRA and full tuning alongside DPO and PPO, multimodal models, a command line and the LLaMA Board no-code web UI. It sits one band below Megatron-LM, which demonstrates frontier-scale parallelism -- a 462B-parameter model across 6,144 H100 GPUs -- that LLaMA-Factory does not claim.

  • https://github.com/hiyouga/LLaMA-Factory recorded 2026-08-09

    "Unified Efficient Fine-Tuning of 100+ LLMs"; README lists LoRA, QLoRA, full-tuning, DPO and PPO, multimodal support and the "LLaMA Board" web UI

Verified 2026-08-09