VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lin, Juyi, Taherin, Amir, Akbari, Arash, Akbari, Arman, Lu, Lei, Chen, Guangyu, Padir, Taskin, Yang, Xiaomeng, Chen, Weiwei, Li, Yiqian, Lin, Xue, Kaeli, David, Zhao, Pu, Wang, Yanzhi
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914072306384896
author Lin, Juyi
Taherin, Amir
Akbari, Arash
Akbari, Arman
Lu, Lei
Chen, Guangyu
Padir, Taskin
Yang, Xiaomeng
Chen, Weiwei
Li, Yiqian
Lin, Xue
Kaeli, David
Zhao, Pu
Wang, Yanzhi
author_facet Lin, Juyi
Taherin, Amir
Akbari, Arash
Akbari, Arman
Lu, Lei
Chen, Guangyu
Padir, Taskin
Yang, Xiaomeng
Chen, Weiwei
Li, Yiqian
Lin, Xue
Kaeli, David
Zhao, Pu
Wang, Yanzhi
contents Recent large-scale Vision Language Action (VLA) models have shown superior performance in robotic manipulation tasks guided by natural language. However, current VLA models suffer from two drawbacks: (i) generation of massive tokens leading to high inference latency and increased training cost, and (ii) insufficient utilization of generated actions resulting in potential performance loss. To address these issues, we develop a training framework to finetune VLA models for generating significantly fewer action tokens with high parallelism, effectively reducing inference latency and training cost. Furthermore, we introduce an inference optimization technique with a novel voting-based ensemble strategy to combine current and previous action predictions, improving the utilization of generated actions and overall performance. Our results demonstrate that we achieve superior performance compared with state-of-the-art VLA models, achieving significantly higher success rates and 39$\times$ faster inference than OpenVLA with 46 Hz throughput on edge platforms, demonstrating practical deployability. The code is available at https://github.com/LukeLIN-web/VOTE.
format Preprint
id arxiv_https___arxiv_org_abs_2507_05116
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting
Lin, Juyi
Taherin, Amir
Akbari, Arash
Akbari, Arman
Lu, Lei
Chen, Guangyu
Padir, Taskin
Yang, Xiaomeng
Chen, Weiwei
Li, Yiqian
Lin, Xue
Kaeli, David
Zhao, Pu
Wang, Yanzhi
Computer Vision and Pattern Recognition
Artificial Intelligence
Robotics
Recent large-scale Vision Language Action (VLA) models have shown superior performance in robotic manipulation tasks guided by natural language. However, current VLA models suffer from two drawbacks: (i) generation of massive tokens leading to high inference latency and increased training cost, and (ii) insufficient utilization of generated actions resulting in potential performance loss. To address these issues, we develop a training framework to finetune VLA models for generating significantly fewer action tokens with high parallelism, effectively reducing inference latency and training cost. Furthermore, we introduce an inference optimization technique with a novel voting-based ensemble strategy to combine current and previous action predictions, improving the utilization of generated actions and overall performance. Our results demonstrate that we achieve superior performance compared with state-of-the-art VLA models, achieving significantly higher success rates and 39$\times$ faster inference than OpenVLA with 46 Hz throughput on edge platforms, demonstrating practical deployability. The code is available at https://github.com/LukeLIN-web/VOTE.
title VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Robotics
url https://arxiv.org/abs/2507.05116