Cong Liu

The massive size of Large Language Models (LLMs) imposes substantial computational and storage burdens, particularly on devices with limited hardware resources. Compared to foundation models, smaller and more specialized models are often more suitable for practical deployment. Existing customizat...