IPL-sharif/KD_Survey

This resource provides a comprehensive overview of Knowledge Distillation (KD), a technique used to make large, complex AI models (like LLMs or Vision-Language Models) run efficiently on devices with limited computing power, such as edge devices. It takes in various KD methods and categorizes them by their sources, schemes, algorithms, and applications across different data types (like text, speech, 3D input) to output a structured understanding of this field. Data scientists, machine learning engineers, and AI researchers working with large neural networks would use this to understand and apply KD effectively.