Flow Generator Matching

TL;DR

This paper introduces Flow Generator Matching (FGM), a novel method that accelerates flow-matching models to one-step generation with theoretical guarantees, achieving state-of-the-art results and efficient distillation for text-to-image tasks.

Contribution

The paper proposes FGM, a new approach that enables one-step sampling in flow-matching models while preserving performance, and demonstrates its effectiveness through benchmarks and distillation.

Findings

Achieved a new FID score of 3.08 on CIFAR10 with one-step FGM.

Successfully distilled Stable Diffusion 3 into a one-step model.

MM-DiT-FGM rivals multi-step models in quality with single-step efficiency.

Abstract

In the realm of Artificial Intelligence Generated Content (AIGC), flow-matching models have emerged as a powerhouse, achieving success due to their robust theoretical underpinnings and solid ability for large-scale generative modeling. These models have demonstrated state-of-the-art performance, but their brilliance comes at a cost. The process of sampling from these models is notoriously demanding on computational resources, as it necessitates the use of multi-step numerical ordinary differential equations (ODEs). Against this backdrop, this paper presents a novel solution with theoretical guarantees in the form of Flow Generator Matching (FGM), an innovative approach designed to accelerate the sampling of flow-matching models into a one-step generation, while maintaining the original performance. On the CIFAR10 unconditional generation benchmark, our one-step FGM model achieves a new record Fr\'echet Inception Distance (FID) score of 3.08 among few-step flow-matching-based models, outperforming original 50-step flow-matching models. Furthermore, we use the FGM to distill the Stable Diffusion 3, a leading text-to-image flow-matching model based on the MM-DiT architecture. The resulting MM-DiT-FGM one-step text-to-image model demonstrates outstanding industry-level performance. When evaluated on the GenEval benchmark, MM-DiT-FGM has delivered remarkable generating qualities, rivaling other multi-step models in light of the efficiency of a single generation step.