LLaMA-Factory
Unified fine-tuning for 100+ LLMs and VLMs with a no-code web UI.
- Star 数
- 75.5K
- Fork 数
- 9.3K
- 未关闭 Issue
- 1.2K
- 最近提交
- 2026-10-10
数据来自 GitHub API · 更新于 2026-10-15
LLaMA-Factory 是什么？
LLaMA-Factory supports SFT, reward modeling, DPO, PPO and more across a wide range of models, with LoRA/QLoRA and a Gradio “LlamaBoard” UI for configuring and launching runs.
主要功能
- 100+ supported models
- No-code LlamaBoard UI
- LoRA, QLoRA and full training
- Preference optimization
系统要求
- 平台
- Linux · Windows · macOS
- 显卡
- NVIDIA recommended
- Python
- 3.9+
如何安装 LLaMA-Factory
根据项目 README 整理。开源项目更新很快，请以 GitHub 页面上的最新说明为准。
From source
- 1Clone and install.
git clone --depth 1 https://github.com/hiyouga/LLaMA-Factory.git cd LLaMA-Factory pip install -e ".[torch,metrics]" - 2Start the web UI.
llamafactory-cli webui
优势
- Broadest model coverage
- No-code UI
不足
- Many options to tune
LLaMA-Factory 替代品对比
| 软件 | Star 数 | 许可证 | 平台 | 显存 |
|---|---|---|---|---|
| LLaMA-Factory Unified fine-tuning for 100+ LLMs and VLMs with a no-code web UI. | 75.5K | Apache-2.0 | — | |
| Unsloth Fine-tune LLMs 2× faster with up to 70% less memory. | 78K | Apache-2.0 | 3 GB+ depending on model size | |
| Axolotl Config-driven LLM fine-tuning for many models and techniques. | 12.6K | Apache-2.0 | — | |
| TRL Hugging Face’s library for post-training LLMs: SFT, DPO, GRPO. | 19.5K | Apache-2.0 | — | |
| Kohya’s GUI GUI for training LoRA, DreamBooth and fine-tunes for SD, SDXL and FLUX. | 12.6K | Apache-2.0 | 8 GB+ for SD1.5 LoRA; 12–24 GB for SDXL/FLUX | |
| OneTrainer One-stop trainer for diffusion models with a desktop UI. | 3.2K | AGPL-3.0 | 8 GB+ |
常见问题
LLaMA-Factory 免费吗？
是的——LLaMA-Factory 是开源软件，采用 Apache-2.0 许可证。商用前请确认软件及模型权重的许可条款。
LLaMA-Factory 支持 Windows、Mac 和 Linux 吗？
LLaMA-Factory 支持：Linux, Windows, macOS。
运行 LLaMA-Factory 需要什么配置？
NVIDIA recommended
如何安装 LLaMA-Factory？
1. Clone and install. 2. Start the web UI.
LLaMA-Factory 有哪些好的替代品？
常见替代品包括 Unsloth, Axolotl, TRL, Kohya’s GUI，详见本页对比表。