metadata

license: apache-2.0
library_name: transformers
pipeline_tag: robotics

🌏 UniVLA

This is the official checkpoint of our RSS 2025 work: Learning to Act Anywhere with Task-centric Latent Actions

Paper: https://arxiv.org/pdf/2505.06111

Code: https://github.com/OpenDriveLab/UniVLA

🔥 Highlights

A recipe towards generalist policy by planning in a unified, embodiment-agnostic action space.
A novel approach for extracting task-centric latent actions from cross-embodiment videos.
A VLA that achieves state-of-the-art results on multiple benchmarks with compute-efficient training.

How to use

This is the UniVLA pre-trained on human videos (we used the Hand-Object Interaction Track dataset of Ego4D). For finetuning on simulation benchmarks or your customized dataset, please visit our official repo.

📝 Citation

If you find our code or models useful in your work, please cite our paper:

@article{bu2025univla,
  title={UniVLA: Learning to Act Anywhere with Task-centric Latent Actions}, 
  author={Qingwen Bu and Yanting Yang and Jisong Cai and Shenyuan Gao and Guanghui Ren and Maoqing Yao and Ping Luo and Hongyang Li},
  journal={arXiv preprint arXiv:2505.06111},
  year={2025}
}