diffusion

[Paper] Multi-Concept Customization of Text-to-Image Diffusion

요약 Text-to-Image Diffusion 모델에 Customize를 해주겠다 이 논문 좀 특이한게 초장부터 사진질임 Figure 1. Given a few images of a new concept, our method augments a pre-trained text-to-image diffusion model, enabling new generations of the concept in unseen contexts 새로운 사진 몇장으로 기존 Pre-trained text-to-image diffusion model 이 기존에 없던 context를 만들어낸 것이랍니다 Furthermore, we propose a method for composing multiple new concepts tog..