Mean Flow Distillation: Robust and Stable Distillation for Flow Matching Models
For practitioners of generative models, MFD provides a stable and robust distillation method that enables real-time inference for flow matching models without sacrificing quality.
Flow Matching models are slow due to iterative ODE sampling. Mean Flow Distillation (MFD) achieves high-fidelity single-step generation by suppressing high-frequency noise and ensuring trajectory consistency, outperforming prior methods on 4D occupancy forecasting and text-to-image tasks.
Flow Matching models have demonstrated strong performance across a wide range of generative tasks. However, their reliance on ODE-based iterative sampling incurs substantial computational overhead in inference, which limits their applicability in real-time scenes. While distillation is a promising solution, existing approaches largely borrow from diffusion-based score matching, often failing to exploit the intrinsic geometric structure of flows and suffering from training instability, high variance, and degraded generation quality. In this paper, we propose Mean Flow Distillation (MFD), a novel distillation framework tailored for flow matching models. We theoretically demonstrate that MFD acts as a temporal low-pass filter, effectively suppressing the high-frequency optimization noise inherent in variational score distillation (VSD) while ensuring global trajectory consistency. We further prove the Mean Flow Matching Theorem, establishing that matching expected average velocities is sufficient for strict distribution alignment. Empirically, on challenging tasks of high-dimensional manifolds including 4D occupancy forecasting and text-to-image generation, MFD achieves state-of-the-art performance, enabling high-fidelity single-step generation.