Metraipolis › archive › dc_ecffe4715ddf

Representation Conditioning for Generation — arXiv:2610.20608

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.20608
License (as archived): CC0-1.0
Archived: 2026-10-15T21:57:34.762Z
SHA-256 of the archived text: 2882a77722665e6a35a090bfd84241e6eeb3c3e231dcc8f70861dcef178fd20d
Stable URL of this page: https://metraipolis.com/documento/dc_ecffe4715ddf

How to cite

Representation Conditioning for Generation — arXiv:2610.20608 — arXiv (en), https://arxiv.org/abs/2610.20608 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ecffe4715ddf (Metraipolis archive dc_ecffe4715ddf, sha256 2882a77722665e6a35a090bfd84241e6eeb3c3e231dcc8f70861dcef178fd20d)

BibTeX

@misc{dcecffe4715ddf,
  title = {Representation Conditioning for Generation — arXiv:2610.20608},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.20608},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ecffe4715ddf; sha256 2882a77722665e6a35a090bfd84241e6eeb3c3e231dcc8f70861dcef178fd20d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20608. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Representation-alignment methods effectively accelerate the diffusion transformer training. REPA, for example, uses a pretrained vision encoder to supervise an intermediate layer through an MLP projection head which is discarded after training. It aims to import semantics from the vision encoder to the diffusion backbone. While this objective is achieved at early layers by back-propagating the alignment loss, there is no guarantee that later layers, supervised by the velocity prediction loss alone, can well access such semantics. We propose a simple solution, representation conditioning (RECON), to improve semantics in later layers. During training, we feed the pretrained vision encoder tokens into later layers; during inference, because we no longer have the clean image, we reuse the MLP head (as a proxy to the vision encoder) and project its output tokens to the later layers. Our method adds no parameters and leads to much faster convergence: RECON is $63\times$ faster than SiT and $25\times$ faster than REPA. Code is available at this https URL

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.20608 — 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 2026-10-15T21:38:29.556Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).