Paper page - Energy-Guided Flow Matching

Source: https://huggingface.co/papers/2608.05811 Published on Aug 7

·

Submitted byhttps://huggingface.co/ysng

gon Aug 19

Abstract

Energy-Guided Flow Matching improves generative quality by progressively revealing high-frequency details through a moving endpoint and adaptive scheduling, reducing training cost and achieving state-of-the-art FID scores.

Pixel-space generative models bypass lossy latent compression, yet necessitate joint learning of global structure and fine-grained details in a high-dimensional space. Standardflow matchinginterpolates noise toward a fixed clean-image endpoint, leaving the spectral evolution to be learned implicitly. In this paper, we introduce Energy-GuidedFlow Matching(EG-FM) that explicitly models a coarse-to-fine generative trajectory by moving endpoint. Specifically, EG-FM replaces the fixed endpoint with aheat-kernel-filtered endpointthat evolves smoothly from low-frequency image to clean image. The fraction of high-frequency signal in moving endpoint is released by an image-specificenergy-guided scheduling, leading to the re-targeting of velocity inflow matching. Our framework requires no adaptation of the backbone and training data, bringing negligible cost on the training and inference stages. In our experiment, EG-FM consistently achieves lower FID on the ImageNetclass-conditional image generationtask at 256 times 256 with fewer epochs, reaching an FID of 1.55 at 200 epochs and 1.45 at 600 epochs. We continue training the generation task on the setting of 512 times 512 resolution, yielding a FID of 1.58 after only 40 high-resolution adaptation epochs. Furthermore, we transfer EG-FM ontext-to-image generationand achieve 0.85 on GenEval score and 83.9 on DPG-Bench. Code is available at https://github.com/ysng123/EG-FM.

View arXiv pageView PDFProject pageGitHub15Add to collection

Get this paper in your agent:

hf papers read 2608\.05811

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper1

#### ysng/EG-FM-ImageNet Unconditional Image Generation• Updated7 minutes ago

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.05811 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.05811 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.