Evo-1: Lightweight Vision-Language-Action Model with Preserved Semantic Alignment

Open in new window