Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots

论文概要

研究领域: CV 作者: Ling Xu, Chuyu Han, Borui Li, Hao Wu, Shiqi Jiang, Ting Cao, Chuanyou Li, Sheng Zhong, Shuai Wang 发布时间: 2026-07-02 arXiv: 2607.02501 分类: cs.RO, cs.CV, cs.OS

中文摘要

具身AI模型现已涵盖视觉-语言-动作（VLA）模型和世界-动作模型（WAM），但实际部署仍因模型特定的Python技术栈、后端假设和机器人端粘合代码而碎片化，尤其在异构边缘设备上。现有推理运行时主要面向请求-响应服务设计，因此无法满足具身部署的运行时契约：闭环控制内的多速率执行、异构硬件上的延迟优先批大小为1的推理，以及超越固定token I/O的可扩展具身接口。本文提出Embodied.cpp，一种面向具身模型的可移植C++推理运行时。基于对代表性VLA模型和WAM的架构分析，Embodied.cpp提取共享执行路径，并将其组织为五层：输入适配器、序列构建器、主干执行、头部插件和部署适配器。该运行时提供模块化多速率执行、延迟优先融合推理，以及可扩展的算子和I/O支持，通过单一后端抽象实现跨异构设备、机器人和模拟器的部署。我们在两个VLA模型（HY-VLA和pi0.5）以及使用LingBot-VA Transformer块的初步WAM基准上评估了Embodied.cpp。VLA部署分别实现了100.0%和91.0%的任务成功率，成功完成闭环执行。WAM基准将块内存从312.2 MiB降至88.1 MiB。这些结果表明，Embodied.cpp在保持高精度的同时提升了部署效率，并适用于多种具身模型架构。

原文摘要

Embodied AI models now span vision-language-action (VLA) models and world-action models (WAMs), but practical deployment remains fragmented across model-specific Python stacks, backend assumptions, and robot-side glue code, especially on heterogeneous edge devices. Existing inference runtimes are designed mainly for request-response serving and therefore do not satisfy the runtime contract of embodied deployment: multi-rate execution inside closed-loop control, latency-first batch-1 inference on heterogeneous hardware, and extensible embodied interfaces beyond fixed token I/O. We present Embodied.cpp, a portable C++ inference runtime for embodied models. Based on an architectural analysis of representative VLA models and WAMs, Embodied.cpp captures a shared execution path and organizes it ...

--- *自动采集于 2026-07-06*

#论文 #arXiv #CV #小凯