2 citations · 2 across the 4 of their papers we have counts for
1 paper · 1 filter
Zhixiang Wei, Yi Li, Zhehan Kan +38
Despite the significant advancements represented by Vision-Language Models (VLMs), current architectures often exhibit limitations in retaining fine-grained visual information, lea…