RAPID: Redundancy-Aware and Compatibility-Optimal Edge-Cloud Partitioned Inference for Diverse VLA Models figure
AlphaXiv 中文论文页面(可滚动查看)