Metraipolis › archive › dc_22dcbc676526
Rethinking Classifier-Free Guidance for TI2I Flow Matching — arXiv:2610.21550
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.21550
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T10:25:13.943Z
- SHA-256 of the archived text
20dcc4b65df2200996c83b04f77f35b3da2085920190ed458b3fd18ecdf529ea- Stable URL of this page
https://metraipolis.com/documento/dc_22dcbc676526
How to cite
Rethinking Classifier-Free Guidance for TI2I Flow Matching — arXiv:2610.21550 — arXiv (en), https://arxiv.org/abs/2610.21550 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_22dcbc676526 (Metraipolis archive dc_22dcbc676526, sha256 20dcc4b65df2200996c83b04f77f35b3da2085920190ed458b3fd18ecdf529ea)
BibTeX
@misc{dc22dcbc676526,
title = {Rethinking Classifier-Free Guidance for TI2I Flow Matching — arXiv:2610.21550},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.21550},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_22dcbc676526; sha256 20dcc4b65df2200996c83b04f77f35b3da2085920190ed458b3fd18ecdf529ea},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21550. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We study Text-and-Image-to-Image (TI2I) flow matching methods for controllable image editing. In a broader generative modeling context, Classifier-Free Guidance (CFG) is established as a key component responsible for high quality outputs. However, we argue that its role in TI2I modeling is poorly understood. Specifically, we make the surprising observation that the CFG mechanism is implicitly harmful for editing in the noisy phase of flow modeling, hindering instruction following and controllability. We propose a simple correction in the early flow steps, enabling a richer alignment-preservation tradeoff at no extra cost. The effectiveness of our approach is demonstrated on both pixel-space and latent TI2I flows. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.21550 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T08:48:02.909Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).