DV-SFT: Direct Vision Supervision for Fine-Grained Visual Understanding 文章

ArXiv CS.CV2026-05-27NEWSen作者: Jianfei Zhao, Feng Zhang, Xin Sun, Chong Feng, Bing Wang, Zhixing Tan
大语言模型
查看原文 →
关系图谱

DV-SFT: Direct Vision Supervision for Fine-Grained Visual Understanding · 相关技术

相关技术

ORMTTEMultimodal Large Language Models (MLLMs)ODELLMlanguage model多模态UCTTokenStanSSESFTReferring expression comprehension (REC)Parts-of-Speech (POS) tagsNext-token predictionNPUMITlarge language modelsICTHISForDiTDecodeCamouflaged object detectionARGANNTable