一作ACL 2026Poster· Natural Language Processing
Investigating Cross-Modal Skill Injection: Scenarios, Methods, and Hyperparameters
Zhiyu Xu · Lean Wang · Yuanxin Liu · Lei Li · Hao Zhou · Fandong Meng · Jie Zhou · Xu Sun
摘要
Vision-Language Models (VLMs) have demonstrated remarkable proficiency in general multi-modal understanding; yet they struggle to efficiently acquire continually evolving domain-specific skills. Conventional approaches to enhancing VLM capabilities, such as Supervised Fine-Tuning (SFT), require e...