Submitted by Polina Fedotova 284 Green-VLA: Staged Vision-Language-Action Model for Generalist Robots Sber Robotics Center 36 7