arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.05964cs.ROcs.AI
收藏

训练时的动作条件化以实现高效的实时分块

Training-Time Action Conditioning for Efficient Real-Time Chunking

Physical Intelligence(物理智能)

Kevin Black, Allen Z. Ren, Michael Equi, Sergey Levine

AI总结:

本研究提出在训练时模拟推理延迟并直接条件化动作前缀,以实现更高效、计算成本更低的实时机器人控制。

摘要:

实时分块(RTC)使视觉-语言-动作模型(VLAs)能够通过异步预测动作分块并利用推理时的图像修复来生成流畅、反应性的机器人轨迹。然而,这种图像修复方法引入了计算开销,增加了推理延迟。在本工作中,我们提出了一种简单的替代方法:在训练时模拟推理延迟,并直接对动作前缀进行条件化,从而消除任何推理时的开销。我们的方法不需要修改模型架构或机器人运行时,仅需用几行额外代码即可实现。在模拟实验中,我们发现训练时的RTC在更高的推理延迟下优于推理时的RTC。在使用$π_{0.6}$ VLA的现实世界实验中,我们展示了训练时的RTC在保持任务性能和速度与推理时的RTC相当的同时,计算成本更低。我们的结果表明,训练时的动作条件化是实时机器人控制中推理时图像修复的实用替代方案。

英文摘要:

Real-time chunking (RTC) enables vision-language-action models (VLAs) to generate smooth, reactive robot trajectories by asynchronously predicting action chunks and conditioning on previously committed actions via inference-time inpainting. However, this inpainting method introduces computational overhead that increases inference latency. In this work, we propose a simple alternative: simulating inference delay at training time and conditioning on action prefixes directly, eliminating any inference-time overhead. Our method requires no modifications to the model architecture or robot runtime, and can be implemented with only a few additional lines of code. In simulated experiments, we find that training-time RTC outperforms inference-time RTC at higher inference delays. In real-world experiments on box building and espresso making tasks with the $π_{0.6}$ VLA, we demonstrate that training-time RTC maintains both task performance and speed parity with inference-time RTC while being computationally cheaper. Our results suggest that training-time action conditioning is a practical drop-in replacement for inference-time inpainting in real-time robot control.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑