Vision Language Models, Efficient Inferencing