发表机构
Max Planck Institute for Intelligent Systems; The Chinese University of Hong Kong; Johannes Kepler University Linz（马克斯·普朗克智能系统研究所; 香港中文大学; 约翰·开普勒林茨大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
WAMJET是一个智能体驾驭工具，通过瓶颈驱动的工作流程和编码智能体自动组合优化技术，为世界动作模型推理实现高达9.95倍的无损加速，同时保持动作质量。
AI 中文摘要
世界动作模型（WAMs）利用预训练的视频基础模型进行机器人操作，但其庞大的主干网络和视频-动作联合预测成本高昂。尽管现有的加速技术提供了多种降低该成本的方法，但针对每个模型和硬件平台选择和组合这些技术需要大量的工程工作。为应对这一瓶颈，我们提出了WAMJET，一个智能体驾驭工具，通过为编码智能体配备可复用的优化指导以及测量和验证工具来加速WAM推理。WAMJET遵循瓶颈驱动的工作流程，智能体对推理进行性能分析，修改目标代码，验证效果，并随着瓶颈的转移迭代地优化加速栈，同时保持动作质量。实验涵盖六个WAM、三个编码智能体和两种GPU架构。WAMJET相对于上游实现实现了高达9.95倍的无损加速。近似和硬件感知优化进一步降低了延迟，同时保持了相当的成功率。结果表明，WAMJET能够为WAM部署生成有效的加速栈。
英文摘要
World Action Models (WAMs) leverage pretrained video foundation models for robot manipulation, but their large backbones and video-action co-prediction are expensive. Although existing acceleration techniques offer many ways to reduce this cost, selecting and composing them requires substantial engineering for each model and hardware platform. To tackle this bottleneck, we present WAMJET, an agentic harness that accelerates WAM inference by equipping coding agents with reusable optimization guidance and measurement and validation tools. WAMJET follows a bottleneck-driven workflow where the agent profiles inference, modifies targeted code, validates effects, and iteratively refines the acceleration stack as bottlenecks shift, while preserving action quality. Experiments span six WAMs, three coding agents, and two GPU architectures. WAMJET achieves up to 9.95x lossless speedup over upstream implementations. Approximation and hardware-aware optimization yield additional latency reductions, with comparable success rates. The results show that WAMJET can produce effective acceleration stacks for WAM deployment.
Comments8 pages, 3 figures, project page: https://liulixinkerry.github.io/WAMJET/index.html
