FlexiDiT: Your Diffusion Transformer Can Easily Generate High-Quality Samples with Less Compute

Sotiris Anagnostidis; Gregor Bachmann; Yeongmin Kim; Jonas Kohler; Markos Georgopoulos; Artsiom Sanakoyeu; Yuming Du; Albert Pumarola; Ali K. Thabet; Edgar Schönfeld

doi:10.1109/cvpr52734.2025.02637

CVPR2025顶会

FlexiDiT: Your Diffusion Transformer Can Easily Generate High-Quality Samples with Less Compute

Sotiris Anagnostidis, Gregor Bachmann, Yeongmin Kim, Jonas Kohler, Markos Georgopoulos, Artsiom Sanakoyeu, Yuming Du, Albert Pumarola, Ali K. Thabet, Edgar Schönfeld

DOI 出版方

2025年份

10顶会引用

摘要

agnostic to input and conditioning modalities. We show how our approach can be readily extended for video generation, where FlexiDiT models generate samples with up to 75% less compute without compromising performance.