StyleDrop: Text-to-Image Synthesis of Any Style

Kihyuk Sohn; Lu Jiang; Jarred Barber; Kimin Lee; Nataniel Ruiz; Dilip Krishnan; Huiwen Chang; Yuanzhen Li; Irfan Essa; Michael Rubinstein; Yuan Hao; Glenn Entis; Irina Blok; Daniel Castro Chin

doi:10.52202/075280-2920

NeurIPS2023顶会

StyleDrop: Text-to-Image Synthesis of Any Style

Kihyuk Sohn, Lu Jiang, Jarred Barber, Kimin Lee, Nataniel Ruiz, Dilip Krishnan, Huiwen Chang, Yuanzhen Li, Irfan Essa, Michael Rubinstein, Yuan Hao, Glenn Entis

DOI 出版方

2023年份

71被引次数

24顶会引用

摘要

Visualization of StyleDrop outputs generated by personalized text-to-image models for 18 different styles. Each model is tuned on a single style reference image, which is shown in the white insert box of each image. The per-style text descriptor is appended to the content text prompt: "A fluffy baby sloth with a knitted hat trying to figure out a laptop, close up". Generated images capture many nuances such as colors, shading, textures and 3D appearance.