평범한 필기장

Paper : https://arxiv.org/abs/2411.16819 Pathways on the Image Manifold: Image Editing via Video GenerationRecent advances in image editing, driven by image diffusion models, have shown remarkable progress. However, significant challenges remain, as these models often struggle to follow complex edit instructions accurately and frequently compromise fidelity byarxiv.orgAbstract기존 문제 기존의 image dif..

Paper : https://arxiv.org/abs/2412.09611 FluxSpace: Disentangled Semantic Editing in Rectified Flow TransformersRectified flow models have emerged as a dominant approach in image generation, showcasing impressive capabilities in high-quality image synthesis. However, despite their effectiveness in visual generation, rectified flow models often struggle with disentanarxiv.orgAbstract기존 문제 Rectifi..

Paper : https://arxiv.org/abs/2412.15205 FlowAR: Scale-wise Autoregressive Image Generation Meets Flow MatchingAutoregressive (AR) modeling has achieved remarkable success in natural language processing by enabling models to generate text with coherence and contextual understanding through next token prediction. Recently, in image generation, VAR proposes scale-wisarxiv.orgAbstract기존 VAR은 다음 두가지..