robotwin_tri_system_joint_self_attention

An OpenWAM-Study checkpoint from the Q1 architecture comparison, run on RoboTwin 2.0 under a fixed recipe so that only the composition rule changes. The baseline is robotwin_dual_system_joint_self_attention.

Tri-System / Joint Self-Attention. Extends the dual layout with a frozen vision-language backbone feeding a separate trainable understanding stream. The joint sequence spans all three streams, with the understanding stream joining as a read-only tail the others attend to.

• Paper: https://arxiv.org/abs/2609.07398
• Code: https://github.com/OpenWAM-Official/OpenWAM
• Project page: https://openwam-official.github.io/

Citation

@article{wang2026openwam,
  title   = {OpenWAM: An Open, Modular Exploration Towards Systematic World-Action Model Pretraining},
  author  = {Yuran Wang and Siqiao Huang and Mingleyang Li and Chenhao Zhang and Jiaqi Liang and Weiyang Jin and Yue Chen and Xuemin Chi and Donghao Zhou and Qize Yu and Yu-Kai Wang and Yuhan Rui and Shenzhe Yao and Zhen Yuan and Zhenhao Shen and Kefei Zhu and Zijie Zhu and Ning Gao and Xiaowei Chi and Guanqi He and Shanghang Zhang and Hao Dong and Lin Shao and Hang Zhao},
  year    = {2026},
  journal = {arXiv preprint arXiv: 2609.07398}
}