arXiv AI
VLA-HWBench: Where the Time Goes When VLA Models Run on Edge Hardware
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv AI
Jun 9
vla.cpp: A Unified Inference Runtime for Vision-Language-Action Models
arXiv:2606. 08094v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) policies are typically shipped as Python/PyTorch stacks that assume a workstation-class GPU, a mismatch for the hardware on which robots actually run.
arXiv AI
Jul 15
Jetson-PI: Towards Onboard Real-Time Robot Control via Foresight-Aligned Asynchronous Inference
arXiv:2607. 12659v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have achieved impressive performance on diverse embodied tasks.
arXiv AI
Aug 18
Efficient Block-Layer Parallel Inference for Vision-Language-Action on Hybrid Architectures
arXiv:2608. 14586v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models are becoming a promising paradigm for autonomous driving, but their deployment on existing vehicle platforms remains difficult because they introduce both high inference latency and strong GPU-side resource pressure.
arXiv AI
Aug 5
PhyAI: Real-Time Physical AI at the Edge, Scalable Rollouts in the Cloud
arXiv:2608. 03682v1 Announce Type: new Abstract: Physical AI policies require inference throughout their lifecycle, including model evaluation, cloud reinforcement learning rollout, edge GPU serving, and onboard deployment.
arXiv Machine Learning
Sep 25
Decoupled Early Exits for Task-Dependent Compute Allocation in Flow-Matching VLAs
The paper introduces a framework for Flow‑Matching Vision‑Language‑Action (VLA) models that allows independent adjustment of backbone depth, action expert depth, and denoising steps. Lightweight Exit Transformers are added at intermediate layers to enable early exits, and a KV Cache synthesis mechanism manages skipped layers so the action expert can exit deeper than the backbone. Experiments on SmolVLA and π0.5 across LIBERO and Meta‑World show that joint tuning of these compute axes reduces latency by 79.2 % and FLOPs by 31.8 %, while improving mean success rate by 5.6 %.
arXiv Machine Learning
Sep 17
VLA-ULAP: Interleaving Cloud VLA Calls with Ultra-Lightweight Local Action Prediction at the Edge
The paper introduces VLA-ULAP, a lightweight local action predictor that interleaves remote vision–language–action (VLA) calls with on‑edge inference. ULAP, with only 7.4 M parameters, predicts action chunks in a single pass using current views, proprioception, and action history, eliminating the need for VLA hidden states or server round‑trips. Experiments on Jetson Orin Nano and simulated benchmarks show that VLA-ULAP can remove 48.8–76.7 % of VLA calls while preserving 95–97.5 % of baseline success, and it outperforms local VLA‑acceleration alternatives in both inference time and energy consumption.