Skip to content

lmms-finetune

Fine-tune multimodal models in one codebase

Visit official site

What lmms-finetune is

The codebase is described as a unified, lightweight framework for integrating and fine-tuning large multimodal models.

What you can do with it

  • Clone the repository into a local environment
  • Set up a conda environment with Python 3.10
  • Install dependencies from requirements.txt

Model notes

  • LLaVA-1.5 and Phi-3-Vision are listed
  • Qwen-VL-Chat and LLaVA-NeXT variants are listed
  • Video and interleaved-image models are supported

Pros & cons

✓ What we like

  • Detailed docs and examples are provided
  • The codebase is described as simple and light

! What to watch out for

  • A conda environment with Python 3.10 is documented
  • Extra libraries such as flash attention are installed as needed

FAQ

Which strategies are named?

The material lists full fine-tuning, LoRA and Q-LoRA.

Which models are supported?

LLaVA-1.5, Phi-3-Vision, Qwen-VL-Chat and LLaVA-NeXT variants are named.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review