Vlaser: Vision-Language-Action Model with Synergistic Embodied Reasoning - View it on GitHub