See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Feng, Yixu, Zhao, Zinan, Ma, Yanxiang, Xia, Chenghao, Du, Chengbin, Wang, Yunke, Xu, Chang
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910227938410496
author Feng, Yixu
Zhao, Zinan
Ma, Yanxiang
Xia, Chenghao
Du, Chengbin
Wang, Yunke
Xu, Chang
author_facet Feng, Yixu
Zhao, Zinan
Ma, Yanxiang
Xia, Chenghao
Du, Chengbin
Wang, Yunke
Xu, Chang
contents Vision-Language-Action (VLA) models have shown remarkable promise in robotics manipulation, yet their high computational cost hinders real-time deployment. Existing token pruning methods suffer from a fundamental trade-off: aggressive compression using pruning inevitably discards critical geometric details like contact points, leading to severe performance degradation. This forces a compromise, limiting the achievable compression rate and thus the potential speedup. We argue that breaking this trade-off requires rethinking compression as a geometry-aware, continuous token resampling in the vision encoder. To this end, we propose the Differentiable Grid Sampler (GridS), a plug-and-play module that performs task-aware, continuous resampling of visual tokens in VLA. By adaptively predicting a minimal set of salient coordinates and extracting features via differentiable interpolation, GridS preserves essential spatial information while achieving drastic compression (with fewer than 10% original visual tokens). Experiments on both LIBERO benchmark and a real robotic platform demonstrate that validating the lowest feasible visual token count reported to date, GridS achieves a 76% reduction in FLOPs with no degradation in the success rate. The code is available at https://github.com/Fediory/Grid-Sampler.
format Preprint
id arxiv_https___arxiv_org_abs_2605_11817
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model
Feng, Yixu
Zhao, Zinan
Ma, Yanxiang
Xia, Chenghao
Du, Chengbin
Wang, Yunke
Xu, Chang
Robotics
Computer Vision and Pattern Recognition
Vision-Language-Action (VLA) models have shown remarkable promise in robotics manipulation, yet their high computational cost hinders real-time deployment. Existing token pruning methods suffer from a fundamental trade-off: aggressive compression using pruning inevitably discards critical geometric details like contact points, leading to severe performance degradation. This forces a compromise, limiting the achievable compression rate and thus the potential speedup. We argue that breaking this trade-off requires rethinking compression as a geometry-aware, continuous token resampling in the vision encoder. To this end, we propose the Differentiable Grid Sampler (GridS), a plug-and-play module that performs task-aware, continuous resampling of visual tokens in VLA. By adaptively predicting a minimal set of salient coordinates and extracting features via differentiable interpolation, GridS preserves essential spatial information while achieving drastic compression (with fewer than 10% original visual tokens). Experiments on both LIBERO benchmark and a real robotic platform demonstrate that validating the lowest feasible visual token count reported to date, GridS achieves a 76% reduction in FLOPs with no degradation in the success rate. The code is available at https://github.com/Fediory/Grid-Sampler.
title See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.11817