Bootstrap Your Generator: Unpaired Visual Editing with Flow Matching

Hugging Face Daily Papers Papers
flow-matching image-editing video-editing unpaired-training generative-models gradient-routing nvidia

Summary

Bootstrap Your Generator (ByG) is a framework for unpaired training of flow matching editing models, leveraging base model knowledge and gradient routing to achieve state-of-the-art results in data-scarce image and video editing tasks.

Modern generative models possess a deep understanding of visual content, yet training them for image editing typically requires massive datasets of paired examples. This limits scalability, especially for video editing where collecting paired data is prohibitively expensive. We propose Bootstrap Your Generator (ByG), a general framework for unpaired training of flow matching editing models. It leverages the base model's knowledge without any external signal. Our approach pairs instruction-following cues extracted from the frozen model with cycle-consistency for structure preservation. To make this tractable, we propose to route gradients from downstream losses over clean predictions to noisy training states. We demonstrate state-of-the-art results on challenging data-scarce image and video editing scenarios. Extensive evaluations and user studies show that our method effectively generalizes to unseen domains and outperforms supervised baselines trained on millions of samples. Analysis reveals that our gradient routing bridges the train-inference gap, and extracting semantic cues from a base model provides a robust training signal that obviates the need for external reward models.
Original Article
View Cached Full Text

Cached at: 06/03/26, 03:38 PM

Paper page - Bootstrap Your Generator: Unpaired Visual Editing with Flow Matching

Source: https://huggingface.co/papers/2606.03911

Abstract

Bootstrap Your Generator framework enables unpaired training of flow matching editing models by leveraging base model knowledge and gradient routing for improved generalization in data-scarce scenarios.

Modern generative models possess a deep understanding of visual content, yet training them for image editing typically requires massive datasets of paired examples. This limits scalability, especially for video editing where collecting paired data is prohibitively expensive. We propose Bootstrap Your Generator (ByG), a general framework forunpaired trainingofflow matchingediting models. It leverages thebase model’s knowledge without any external signal. Our approach pairsinstruction-following cuesextracted from the frozen model withcycle-consistencyfor structure preservation. To make this tractable, we propose to route gradients fromdownstream lossesover clean predictions to noisy training states. We demonstrate state-of-the-art results on challenging data-scarce image and video editing scenarios. Extensive evaluations and user studies show that our method effectively generalizes to unseen domains and outperforms supervised baselines trained on millions of samples. Analysis reveals that ourgradient routingbridges thetrain-inference gap, and extractingsemantic cuesfrom abase modelprovides a robust training signal that obviates the need for external reward models.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2606\.03911

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

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.03911 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.03911 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.03911 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.

Similar Articles

Follow the Mean: Reference-Guided Flow Matching

Hugging Face Daily Papers

This paper introduces a method for controllable generation in flow matching by adjusting the conditional endpoint mean using a reference set, offering both training-free and semi-parametric guidance for style and content control.

LeapAlign: Post-Training Flow Matching Models at Any Generation Step by Building Two-Step Trajectories

Hugging Face Daily Papers

LeapAlign is a post-training method that improves flow matching model alignment with human preferences by reducing computational costs through two-step trajectory shortcuts while enabling stable gradient propagation to early generation steps. The method outperforms state-of-the-art approaches when fine-tuning Flux models across various image quality and text-alignment metrics.