Shortcutting Pre-trained Flow Matching Diffusion Models is Almost Free Lunch

TL;DR

This paper introduces a highly efficient post-training method for transforming large pre-trained flow matching diffusion models into few-step samplers using velocity field self-distillation, significantly reducing computational costs.

Contribution

The authors propose a novel velocity field self-distillation technique that enables aggressive shortcutting in flow matching models without retraining, improving efficiency and enabling few-shot distillation.

Findings

Achieved 3-step Flux sampling in less than one A100 day.

Enabled few-shot distillation with as few as 10 text-image pairs.

Produced state-of-the-art performance at minimal cost.

Abstract

We present an ultra-efficient post-training method for shortcutting large-scale pre-trained flow matching diffusion models into efficient few-step samplers, enabled by novel velocity field self-distillation. While shortcutting in flow matching, originally introduced by shortcut models, offers flexible trajectory-skipping capabilities, it requires a specialized step-size embedding incompatible with existing models unless retraining from scratch$\unicode{x2013}$a process nearly as costly as pretraining itself. Our key contribution is thus imparting a more aggressive shortcut mechanism to standard flow matching models (e.g., Flux), leveraging a unique distillation principle that obviates the need for step-size embedding. Working on the velocity field rather than sample space and learning rapidly from self-guided distillation in an online manner, our approach trains efficiently, e.g., producing a 3-step Flux less than one A100 day. Beyond distillation, our method can be incorporated into the pretraining stage itself, yielding models that inherently learn efficient, few-step flows without compromising quality. This capability also enables, to our knowledge, the first few-shot distillation method (e.g., 10 text-image pairs) for dozen-billion-parameter diffusion models, delivering state-of-the-art performance at almost free cost.