LLaMA-Factory — файн-тюнинг языковых моделей

★ 7.7 · ai-tooling

llama-factory is a Claude Code skill that provides expert guidance for fine-tuning large language models using LLaMA-Factory. It supports 100+ models including Llama, Qwen, and Gemma, with QLoRA quantization in 2/3/4/5/6/8-bit modes and multimodal fine-tuning via a no-code WebUI. The skill bundles official LLaMA-Factory documentation in a references/ directory, covering getting started, advanced configurations, and image-related workflows. Core dependencies include llmtuner, torch, transformers, peft, accelerate, and gradio. It is aimed at developers and researchers who want to fine-tune LLMs for domain-specific tasks without deep infrastructure expertise.