[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-multimodal-llm":3},{"tag":4,"articles":10},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":9},"440ca6d5-87cb-4c38-979b-a78c173102b7","multimodal llm","multimodal-llm",0,null,[11],{"id":12,"slug":13,"title":14,"summary":15,"category":16,"image_url":17,"cover_image":17,"language":18,"created_at":19},"d2789ba9-bfb3-43bc-a14f-5212834da5a8","mllms-subject-driven-image-generation-en","MLLMs for cleaner subject-driven image generation","This paper uses MLLMs plus VAE identity conditioning to improve subject-driven image generation and reduce copy-paste artifacts.","research","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1779776770975-82hd.png","en","2026-05-26T06:25:40.47126+00:00"]