Same Path, Less Time: Training-Free Fast Execution of Vision-Language-Action Models under Inference Latency
Same Path, Less Time: Training-Free Fast Execution of Vision-Language-Action Models under Inference Latency
This research presents Path-preserving Time Scaling (PTS), a training-free execution layer designed for frozen vision-language-action (VLA) models to mitigate the time wasted during inference and slow demonstration speeds.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
model-suite combinations tested
6
View original evidence
under injected inference latency, PTS raises task throughput over the strongest deployable baseline by 33% for pi0.5 and 104% for SmolVLA while matching or improving the success rate, is best on all six model-suite combinations, and exceeds zero-latency 1x execution. Against asynchronous 1x execution with the same one-step decoder, the time layer alone adds 28% and 38%. On LIBERO-Dynamic it improves on stop-and-wait
Open source S5inference latency condition
1
View original evidence
arXiv:2611.01390v1 Announce Type: new Abstract: Vision-language-action (VLA) models predict chunks of future actions, yet deployed robots waste time in two ways that are independent of model capability: they halt while the next chunk is inferred, and they move at the speed of slow, cautious demonstrations. Existing remedies remove one loss each, and stacking them fails, because asynchrony and speed-up each change the
Open source S2
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
under injected inference latency, PTS raises task throughput over the strongest deployable baseline by 33% for pi0.5 and 104% for SmolVLA while matching or improving the success rate, is best on all six model-suite combinations, and exceeds zero-latency 1x execution. Against asynchronous 1x execution with the same one-step decoder, the time layer alone adds 28% and 38%. On LIBERO-Dynamic it improves on stop-and-wait
Open source S5
execution by 38% and is comparable to asynchronous 1x execution.
Open source S6
ss consumed during inference. A time layer sums the planned increments, keeps the arm at 1x while the gripper moves, and caps the rate by the actuator's linear range and by chunk feasibility under latency. We show that misaligned entry and stochastic sampling add a bias and a variance to the continuation error, and derive a break-even success rate that states when speeding up pays off. On LIBERO with two frozen VLAs
Open source S4
arXiv:2611.01390v1 Announce Type: new Abstract: Vision-language-action (VLA) models predict chunks of future actions, yet deployed robots waste time in two ways that are independent of model capability: they halt while the next chunk is inferred, and they move at the speed of slow, cautious demonstrations. Existing remedies remove one loss each, and stacking them fails, because asynchrony and speed-up each change the
Open source S2
Source：arXiv Robotics — research abstracts · arxiv.org