UnifoLM-VLA-0
- AI model
- Humanoids and robot learning
A vision-language-action model for general-purpose humanoid robot manipulation, continued from robot manipulation data to add physical common sense to a vision-language model.
- No license
- Non-commercial weights
Made by Unitree Robotics, China (checked on github.com/unitreerobotics (opens in a new tab)).
Source on GitHub: github.com/unitreerobotics/unifolm-vla (opens in a new tab). Facts read at commit ff6c39a (opens in a new tab).
Licenses
- Code
- No license
No license file at the repository's root at commit ff6c39a (opens in a new tab), on 16 Oct 2026.
- Weights
- CC-BY-NC-SA-4.0 (opens in a new tab)
Read on huggingface.co/unitreerobotics/UnifoLM-VLA-Base (opens in a new tab).
- Commercial use
- No
The weights are under CC BY-NC-SA 4.0 (non-commercial). The code has no license file; its pyproject.toml only declares BSD-3-Clause.
Labels
- No licenseNo license file at the repository's root: the source is available, but all rights are reserved.
- Non-commercial weightsThe weights license (CC-BY-NC-SA-4.0) does not allow commercial use.
What it does: Robotheca's summary
- Continued pre-training of a vision-language model on robot manipulation data for spatial and geometric understanding.
- Uses full dynamics prediction data to generalize across diverse manipulation tasks.
- Real-robot validation reports one policy completing 12 categories of complex manipulation tasks.
- Released with training and inference code and checkpoints for a base model and a LIBERO-finetuned model.
- Supports fine-tuning on RLDS datasets converted from LeRobot V2.1 format, with twelve open-source Unitree G1 datasets used in experiments.
- Includes a LIBERO simulation evaluation pipeline and a server-client real-robot inference setup.
Written by Robotheca from the project's README (opens in a new tab) at commit ff6c39a.
Getting started
- Difficulty
- Advanced
- How to start
- The project's README (opens in a new tab)
Maintenance
- Status
- Maintained: a commit in the last 12 months
- Newest commit
- 29 Jan 2026 on main
- Latest release
- None published on GitHub
Checked on GitHub on 31 Oct 2026.
Links
Robots that can run it
No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.
Robotheca lists AI models as a reference: nothing is hosted or run here, and every fact links to where it was read.