Bi-VLA: Vision-Language-Action Model-Based System for Bimanual Robotic Dexterous Manipulations