hiyouga/LlamaFactory

Health v2 · maintenance only

Adoption not part of the score

75500 stars · 9254 forks observed · 2026-10-16

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-21, confidence 0.97

LlamaFactory is an open‑source Python platform that streamlines fine‑tuning, quantization, and inference of over 100 LLM and VLM models. It provides a CLI and a Web UI, supporting LoRA, QLoRA, RLHF, and various acceleration back‑ends.

Use cases

• Fine‑tune a LLaMA or Mistral model on a custom text dataset without writing code.
• Apply LoRA or QLoRA adapters to adapt a large language model for a specific downstream task.
• Quantize a pretrained model to 2‑8 bits for deployment on limited‑resource hardware.
• Run inference with the vLLM or Transformers backend for high‑throughput generation.
• Train multimodal models (e.g., LLaVA) that combine text and images.
• Monitor training experiments using TensorBoard, Wandb, or MLflow.

When to choose

• You need a unified, easy‑to‑use solution for fine‑tuning many different LLM families.
• You want built‑in support for LoRA, QLoRA, RLHF, and other advanced fine‑tuning algorithms.
• You prefer a no‑code or low‑code workflow with both CLI and Web UI options.
• You require integration with popular experiment tracking tools and GPU acceleration kernels.

When to avoid

• You need a production‑grade serving system with autoscaling and high‑availability orchestration.
• Your project relies on non‑Python ecosystems or requires deep custom training loop modifications.
• You need real‑time inference at massive scale beyond what a single‑node or modest multi‑node setup can provide.

Facets

application · maturity active

llm-inference llm-training cli artificial-intelligence machine-learning large-language-models deep-learning python windows cloud self-hosted cli fine-tuning qlora lora quantization rlhf multimodal distributed-training accelerators tensorboard wandb web-ui linux macos docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="hiyouga/LlamaFactory")