Metraipolis › archive › dc_0f31cd50889e

LoopVAE: Recurrent Depth Across Scales for Visual Tokenization — arXiv:2609.11516

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.11516
License (as archived)
CC0-1.0
Archived
2026-10-16T06:35:43.165Z
SHA-256 of the archived text
2c78350c594494537d09702ab062ed92571af69f3d7594b21561bf63f315c5ac
Stable URL of this page
https://metraipolis.com/documento/dc_0f31cd50889e

How to cite

LoopVAE: Recurrent Depth Across Scales for Visual Tokenization — arXiv:2609.11516 — arXiv (en), https://arxiv.org/abs/2609.11516 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0f31cd50889e (Metraipolis archive dc_0f31cd50889e, sha256 2c78350c594494537d09702ab062ed92571af69f3d7594b21561bf63f315c5ac)
BibTeX
@misc{dc0f31cd50889e,
  title = {LoopVAE: Recurrent Depth Across Scales for Visual Tokenization — arXiv:2609.11516},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.11516},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0f31cd50889e; sha256 2c78350c594494537d09702ab062ed92571af69f3d7594b21561bf63f315c5ac},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.11516. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Hierarchical visual tokenizers typically allocate different processing blocks to different spatial scales. We ask how much of this computation can use the same parameters. LoopVAE reuses a scale- and loop-conditioned core within and across scales, while keeping resolution-changing transitions independent. A four-block core executes 28 block applications per encoder or decoder. On ImageNet-256, the 29M-parameter convolutional model reaches 0.28 rFID and 32.54 dB PSNR under an approximately 30-epoch two-stage training budget, using approximately 65% fewer parameters than the 84M reference VAEs. A non-adversarial Transformer ablation with the same execution graph finds competitive PSNR and SSIM under global sharing, although unshared blocks improve LPIPS. Targeted loop interventions show that completing the trained recurrence improves reconstruction and that even small feature updates can have substantial downstream effects. Truncation also exposes output-range errors, distinguishing useful recurrent computation from reliable early exit. Runtime profiling reveals the execution tradeoff: fewer stored weights require more arithmetic and longer runtime in the tested configurations. With convolutional and Transformer operators and single- or multi-resolution latent interfaces, LoopVAE establishes recurrent depth across scales as a parameter-sharing design axis for visual tokenization.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-11 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.11516 — 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 2026-10-16T06:06:28.792Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).