热点
"视觉-文本关联" 相关文章
Inverse-LLaVA: Eliminating Alignment Pre-training Through Text-to-Vision Mapping
cs.AI updates on arXiv.org 2025-08-19T04:21:26.000000Z