Text this: Semantic Scene Completion in Autonomous Driving: A Two-Stream Multi-Vehicle Collaboration Approach