EaqVLA: Encoding-aligned Quantization for Vision-Language-Action Models

TL;DR

EaqVLA introduces an encoding-aligned quantization framework for Vision-Language-Action models, effectively reducing memory and computation costs while maintaining performance, by addressing token misalignment issues.

Contribution

The paper presents a novel encoding-aligned quantization method tailored for VLA models, including an analysis of misalignment and a mixed precision approach to improve quantization performance.

Findings

Achieves minimal quantization loss in end-to-end action control

Provides xxx times acceleration over existing methods

Outperforms previous quantization techniques in VLA models

Abstract

With the development of Embodied Artificial intelligence, the end-to-end control policy such as Vision-Language-Action (VLA) model has become the mainstream. Existing VLA models faces expensive computing/storage cost, which need to be optimized. Quantization is considered as the most effective method which can not only reduce the memory cost but also achieve computation acceleration. However, we find the token alignment of VLA models hinders the application of existing quantization methods. To address this, we proposed an optimized framework called EaqVLA, which apply encoding-aligned quantization to VLA models. Specifically, we propose an complete analysis method to find the misalignment in various granularity. Based on the analysis results, we propose a mixed precision quantization with the awareness of encoding alignment. Experiments shows that the porposed EaqVLA achieves better quantization performance (with the minimal quantization loss for end-to-end action control and xxx times acceleration) than existing quantization methods.