DA-PTQ: Drift-Aware Post-Training Quantization for Efficient Vision-Language-Action Models

TL;DR

DA-PTQ is a novel drift-aware post-training quantization method that reduces kinematic drift in vision-language-action models, enabling efficient deployment on resource-constrained robots without significant performance loss.

Contribution

It introduces a drift-aware optimization framework with cross-space compensation and mixed-precision allocation to improve quantization of VLAs for embodied AI.

Findings

DA-PTQ significantly reduces kinematic drift in VLAs.

It achieves comparable performance to full-precision models at low bit-widths.

The method enables practical deployment of VLAs on resource-limited robotic platforms.

Abstract

Vision-Language-Action models (VLAs) have demonstrated strong potential for embodied AI, yet their deployment on resource-limited robots remains challenging due to high memory and computational demands. While Post-Training Quantization (PTQ) provides an efficient solution, directly applying PTQ to VLAs often results in severe performance degradation during sequential control. We identify temporal error accumulation as a key factor, where quantization perturbations at the vision-language-to-action interface are progressively amplified, leading to kinematic drift in executed trajectories. To address this issue, we propose Drift-Aware Post-Training Quantization (DA-PTQ), which formulates quantization as a drift-aware optimization problem over sequential decision processes. DA-PTQ consists of two components: (1) Cross-Space Representation Compensation, which mitigates structured distortions between multimodal representations and action space to improve action consistency, and (2) Motion-Driven Mixed-Precision Allocation, which assigns bit-widths by minimizing trajectory-level motion errors. Extensive experiments show that DA-PTQ significantly reduces kinematic drift and achieves comparable performance to full-precision models under low-bit settings, enabling practical deployment of VLAs on resource-limited robotic platforms.