FASTer: Toward Efficient Autoregressive Vision Language Action Modeling via Neural Action Tokenization

TL;DR

FASTer introduces a neural action tokenizer and an efficient autoregressive framework for vision-language-action models, significantly improving inference speed and task performance in robotic manipulation.

Contribution

The paper presents FASTer, a novel unified framework combining a learnable tokenizer with autoregressive policies, enhancing efficiency and generalization in robot learning tasks.

Findings

FASTerVQ achieves high-quality action chunk encoding with high compression.

FASTerVLA outperforms previous models in inference speed and task success rates.

Extensive experiments validate FASTer's superior generalization and efficiency.

Abstract

Autoregressive vision-language-action (VLA) models have recently demonstrated strong capabilities in robotic manipulation. However, their core process of action tokenization often involves a trade-off between reconstruction fidelity and inference efficiency. We introduce FASTer, a unified framework for efficient and generalizable robot learning that integrates a learnable tokenizer with an autoregressive policy built upon it. FASTerVQ encodes action chunks as single-channel images, capturing global spatio-temporal dependencies while maintaining a high compression ratio. FASTerVLA builds on this tokenizer with block-wise autoregressive decoding and a lightweight action expert, achieving both faster inference and higher task performance. Extensive experiments across simulated and real-world benchmarks show that FASTerVQ delivers superior reconstruction quality, high token utilization, and strong cross-task and cross-embodiment generalization, while FASTerVLA further improves overall capability, surpassing previous state-of-the-art VLA models in both inference speed and task performance.