Skip to content
RobothecaBeta
Log in

UnifoLM-WLA-1.0 and UnifoLM-ER-1

  • AI model
  • Humanoids and robot learning

Unitree's 6B-parameter humanoid robot foundation model that drives one policy across 64 desktop and whole-body manipulation tasks, with an action-expert training and fine-tuning toolkit.

Made by Unitree Robotics, China (checked on github.com/unitreerobotics (opens in a new tab)).

Source on GitHub: github.com/unitreerobotics/unifolm-wla (opens in a new tab). Facts read at commit 0a1aa87 (opens in a new tab).

Licenses

Code
Apache-2.0 (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (Apache-2.0) at commit 0a1aa87 (opens in a new tab), on 16 Oct 2026.

Commercial use
Yes

What it does

  • A single model coordinates 64 tasks spanning desktop manipulation and whole-body manipulation.
  • Trained on about 2,500 hours of real-robot data, with support for two-finger grippers and multiple five-finger dexterous hands.
  • Built on large-scale multimodal perception data and interaction-centric world modeling for spatial understanding.
  • Provides UnifoLM-ER-1, UnifoLM-ER-Flow and UnifoLM-WLA-1.0-Base model weights on Hugging Face.
  • Includes code for training action experts, fine-tuning from the Base model and LoRA fine-tuning.
  • Documentation covers action, state and statistics processing plus how to train and evaluate an action expert.

Written by Robotheca from the project's README (opens in a new tab) at commit 0a1aa87.

Getting started

Difficulty
Advanced

Maintenance

Status
Active: a commit in the last 90 days
Newest commit
29 Sept 2026 on main
Latest release
None published on GitHub

Checked on GitHub on 16 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.