DyQ-VLA: Temporal-Dynamic-Aware Quantization for Embodied Vision-Language-Action Models

TL;DR

DyQ-VLA introduces a dynamic quantization method for embodied vision-language-action models, reducing memory and computational costs while maintaining high performance through real-time sensitivity-aware bit allocation.

Contribution

It presents a novel dynamic quantization framework that adapts bit-widths in real-time based on sensitivity, addressing limitations of static quantization in VLAs.

Findings

Reduces memory footprint to 30.9% of original

Maintains 99.5% of original performance

Achieves up to 1.43x real-world speedup

Abstract

Vision-Language-Action (VLA) models are dominant in embodied intelligence but are constrained by inference overheads. While model quantization alleviates these bottlenecks for edge deployment, static quantization approaches remain suboptimal for VLAs due to two critical challenges: (1) Temporal-dynamic sensitivity, where fixed precision wastes resources by ignoring stage-varying error tolerances; and (2) Real-time allocation, where identifying real-time sensitivity to guide bit allocation remains unsolved. To address these challenges, we propose DyQ-VLA, a dynamic quantization framework for VLAs. Specifically, a sensitivity-aware switching strategy leverages real-time kinematic proxies to trigger the bit-width switch, while a kinematic-guided module dynamically allocates the optimal bit-width. Experiments show that DyQ-VLA requires only 30.9% of the original memory footprint while maintaining 99.5% of its original performance, achieving 1.49x simulation and up to 1.43x real-world speedups.