SARA-RT: Scaling up Robotics Transformers with Self-Adaptive Robust Attention

TL;DR

SARA-RT introduces a novel fine-tuning method called up-training that efficiently converts large, quadratic-time Robotics Transformers into linear-attention models, enabling faster on-robot deployment without sacrificing performance.

Contribution

The paper proposes up-training, a new fine-tuning approach that transforms pre-trained Robotics Transformers into efficient linear-attention models for practical deployment.

Findings

Speeds up RT-2 vision-language-action models

Accelerates Point Cloud Transformer policies

Maintains high quality after conversion

Abstract

We present Self-Adaptive Robust Attention for Robotics Transformers (SARA-RT): a new paradigm for addressing the emerging challenge of scaling up Robotics Transformers (RT) for on-robot deployment. SARA-RT relies on the new method of fine-tuning proposed by us, called up-training. It converts pre-trained or already fine-tuned Transformer-based robotic policies of quadratic time complexity (including massive billion-parameter vision-language-action models or VLAs), into their efficient linear-attention counterparts maintaining high quality. We demonstrate the effectiveness of SARA-RT by speeding up: (a) the class of recently introduced RT-2 models, the first VLA robotic policies pre-trained on internet-scale data, as well as (b) Point Cloud Transformer (PCT) robotic policies operating on large point clouds. We complement our results with the rigorous mathematical analysis providing deeper insight into the phenomenon of SARA.