InSight-o3: Empowering Multimodal Foundation Models with Generalized Visual Search
Paper • 2512.18745 • Published • 12
InSight-o3
Empowering Multimodal Foundation Models with Generalized Visual Search
-
m-Just/O3-Bench
Viewer • Updated • 345 • 307 • 16Note Can your AI agent truly "think with images"? Test it out on O3-Bench!
m-Just/InSight-o3-vS
Image-Text-to-Text • 8B • Updated • 3Note This is the vSearcher model introduced in our work.
m-Just/VisCoT_VStar_Collage
Viewer • Updated • 15.3k • 131 • 2Note In-loop RL training data for vSearcher.
m-Just/InfoVQA_RegionLocalization
Viewer • Updated • 10.2k • 41 • 1Note Out-of-loop RL training data for vSearcher.