UnifoLM-VLA-0 (and UnifoLM-WMA-0)
HF repos for UnifoLM-VLA-Base created 2026-01-28 (exact announcement day not verified). VLA-0 license CC BY-NC-SA 4.0 (non-commercial); WMA-0 Apache-2.0. Superseded by UnifoLM-WLA-1.0 (2026-09). Unitree also publishes ~200 G1 teleoperation datasets under huggingface.co/unitreerobotics.
- Input
- image, text
- Output
- action
- License
- cc-by-nc-sa-4.0
- Verified
- 2026-09-29
How to call it
| Provider | Model id | Endpoint / URL | Docs |
|---|---|---|---|
| Hugging Face | unitreerobotics/UnifoLM-VLA-Base | huggingface.co/collections/unitreerobotics/unifolm-vla-0 | — |
| Hugging Face (world-model-action) | unitreerobotics/UnifoLM-WMA-0-Base | huggingface.co/unitreerobotics/UnifoLM-WMA-0-Base | — |
| GitHub | — | github.com/unitreerobotics/unifolm-vla | — |
Notable capabilities (2)
- Open VLA for general-purpose humanoid manipulation: Continued pretraining of a VLM (UnifoLM-VLM-Base, Qwen2.5-VL based) on robot manipulation data to turn it into an 'embodied brain'; variants fine-tuned on Unitree open datasets and LIBERO. source
- World-model-action architecture (WMA-0): UnifoLM-WMA-0 (Sept 2025, Apache-2.0) pairs a world model that predicts future interactions (usable as a simulator) with action generation; Base and Dual variants on HF. source