Home /Research /HoloMotion-1 Technical Report
LOCOMOTION

HoloMotion-1 Technical Report

Maiyue Chen, Kaihui Wang, Bo Zhang, Xihan Ma, Zhiyuan Yang, Yi Ren, Qijun Huang, Zihao Zhu, Yucheng Wang, Zhizhong Su

Year
2026
Access
Open access

Abstract

In this report, we present HoloMotion-1, a humanoid motion foundation model for zero-shot whole-body motion tracking. A key innovation of HoloMotion-1 is to scale control-policy training with a large-scale hybrid motion corpus, where video-reconstructed motions from in-the-wild videos provide the dominant source of motion diversity, while curated motion-capture and in-house motion data provide higher-fidelity supervision and deployment-oriented coverage. This data regime enables HoloMotion-1 to move beyond conventional MoCap-only training and exposes the policy to substantially broader behaviors, capture conditions, and motion styles. Learning from such heterogeneous data introduces new challenges, including reconstruction noise, source-domain mismatch, uneven motion quality, and the need for temporal modeling under large behavioral variation. To address these challenges, HoloMotion-1 integrates large-capacity temporal modeling, a sparsely activated Mixture-of-Experts Transformer with KV-cache inference for real-time control, and a sequence-level training strategy that improves learning efficiency on extended motion sequences. Extensive experiments on multiple unseen motion benchmarks show that HoloMotion-1 generalizes robustly across diverse motion types and capture conditions, significantly improves tracking accuracy over prior methods, and transfers directly to a real humanoid robot without task-specific fine-tuning.

Keywords

humanoid motionzero-shot trackingmixture-of-expertsmotion corpuswhole-body control

Related papers

Browse all LOCOMOTION papers