Imagine Flash: Accelerating Emu Diffusion Models with Backward Distillation

arXiv:2405.05224v1 Announce Type: new
Abstract: Diffusion models are a powerful generative framework, but come with expensive inference. Existing acceleration methods often compromise image quality or fail under complex conditioning when operating in an extremely low-step regime. In this work, we propose a novel distillation framework tailored to enable high-fidelity, diverse sample generation using just one to three steps. Our approach comprises three key components: (i) Backward Distillation, which mitigates training-inference discrepancies by calibrating the student on its own backward trajectory; …

abstract arxiv cs.cv diffusion diffusion models distillation diverse emu fidelity flash framework generative image imagine inference low novel quality sample type work