Metraipolis › archive › dc_6930ff964ab4

Diffu-LoRA: A Novel Low-Rank Adaptation for Personalized Diffusion Models — arXiv:2610.10550

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.10550
License (as archived)
CC0-1.0
Archived
2026-10-09T06:03:59.771Z
SHA-256 of the archived text
1756e5d38096687ef06e88820402c2345e3db73a94fb7b22ba37a28a9417c661
Stable URL of this page
https://metraipolis.com/documento/dc_6930ff964ab4

How to cite

Diffu-LoRA: A Novel Low-Rank Adaptation for Personalized Diffusion Models — arXiv:2610.10550 — arXiv (en), https://arxiv.org/abs/2610.10550 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6930ff964ab4 (Metraipolis archive dc_6930ff964ab4, sha256 1756e5d38096687ef06e88820402c2345e3db73a94fb7b22ba37a28a9417c661)
BibTeX
@misc{dc6930ff964ab4,
  title = {Diffu-LoRA: A Novel Low-Rank Adaptation for Personalized Diffusion Models — arXiv:2610.10550},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.10550},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6930ff964ab4; sha256 1756e5d38096687ef06e88820402c2345e3db73a94fb7b22ba37a28a9417c661},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.10550. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Personalizing text-to-image diffusion models from a few reference images requires preserving subject identity while following prompts that describe new contexts. Full-model fine-tuning is parameter-intensive, whereas low-rank adaptation (LoRA) reduces the number of trainable parameters but leaves open how adaptation capacity should be distributed across layers. We introduce Diffu-LoRA, a parameter-efficient method that learns this allocation through gated low-rank adaptation. Diffu-LoRA inserts trainable low-rank components into the linear layers of Transformer blocks and assigns a learnable gate to each component. Bilevel optimization updates the adaptation weights and gate parameters on separate data splits, while progressive pruning removes components with the lowest gate values to meet a prescribed rank budget. This procedure allocates adaptation capacity nonuniformly across layers while keeping the pretrained backbone frozen. Experiments with Stable Diffusion on subjects from DreamBooth and additional collected datasets show improved overall subject fidelity and prompt alignment relative to the evaluated fine-tuning baselines. Ablation studies examine the contributions of bilevel optimization, progressive pruning, and adapter placement. These results support learned rank allocation as a practical approach to parameter-efficient diffusion model personalization.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.10550 — 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 2026-10-09T05:47:15.087Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).