This repository contains the official implementation and benchmark for our paper "In-depth Robustness Analysis for Vision-Language-Action Models". We systematically expose the hidden vulnerabilities of contemporary VLA models through comprehensive robustness evaluation across seven perturbation⦠See the full description on the dataset page:
https://huggingface.co/datasets/Sylvest/LIBERO-plus.