Plain-AyI
← All projects
AI models & research

VLM-R1

Research models and training code that teach image-understanding AI to reason, using reinforcement learning.

Research-backedNew this week

What it does

VLM-R1 is a research project on vision-language models, AI that understands both pictures and text. It applies the training approach made popular by DeepSeek-R1, where the model learns by being rewarded for good answers (reinforcement learning).

The team trained models to find an object in a picture from a written description, to spot objects by name, and to solve maths problems. They report that this approach handled unfamiliar examples better than standard fine-tuning.

The repository includes training code, released models and online demos. It supports popular open models such as Qwen-VL and InternVL.

Who it's for

AI researchers and engineers working on models that reason about images.

Can I use it without coding?

Developers only. Training needs code and powerful hardware, though the online demos can be tried in a browser. New to this? Start here.

Can I run it?

Cost
FreeNothing to pay or sign up for.
Computer
Works best on a powerful computer or one with a graphics card (GPU)
Privacy
Runs on your own device: your data stays with you
API key
Not needed

Try it

  1. Try the online demos linked at the top of the README to see what the models can do.

Using Claude Code, Cursor or another AI coding assistant? Paste the prompt and it will walk you through the setup.New to this?

Before you install

  • Only download from the official GitHub page. Copies with similar names exist for many popular projects.
  • License (Apache-2.0): Permissive: free to use, change and share, including for business.
  • More on staying safe

Where people are talking about it