Deep learning models have transformed fields from computer vision to natural language processing, yet their high computational and energy demands pose critical challenges for widespread deployment.