Metraipolis › archive › dc_9aa40cc40c6d
ConvRoute-ViT: Convolutional Token Routing for Efficient Vision Transformers — arXiv:2610.38284
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.38284
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-01T02:40:22.603Z
- SHA-256 of the archived text
ffe3e2a12491f0801e345f080afd092573e0fc1a58330910e3dc952ae4819d01- Stable URL of this page
https://metraipolis.com/documento/dc_9aa40cc40c6d
How to cite
ConvRoute-ViT: Convolutional Token Routing for Efficient Vision Transformers — arXiv:2610.38284 — arXiv (en), https://arxiv.org/abs/2610.38284 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9aa40cc40c6d (Metraipolis archive dc_9aa40cc40c6d, sha256 ffe3e2a12491f0801e345f080afd092573e0fc1a58330910e3dc952ae4819d01)
BibTeX
@misc{dc9aa40cc40c6d,
title = {ConvRoute-ViT: Convolutional Token Routing for Efficient Vision Transformers — arXiv:2610.38284},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.38284},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9aa40cc40c6d; sha256 ffe3e2a12491f0801e345f080afd092573e0fc1a58330910e3dc952ae4819d01},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.38284. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Vision Transformers (ViTs) achieve strong image recognition performance, but their self-attention increases quadratically with the number of input tokens and makes them computationally expensive. The standard implementation of ViT models treats all patch tokens with equal importance across every layer of the model. This uniform treatment introduces redundant computation and may reduce the model's focus on the most discriminative object regions. In this paper, we propose ConvRoute-ViT, a convolutional token routing mechanism for efficient vision transformers that performs token reduction through convolutional routing. Our model preserves the most relevant patch tokens according to class-token attention at selected layers, while the remaining tokens are spatially reorganized and compressed into a compact set of convolutional summary tokens. This design reduces the sequence length without completely removing the contextual information of the less informative tokens. By applying the routing process hierarchically, the proposed framework constructs compact and discriminative representations for deeper layers. Experiments on CIFAR-100 and TinyImageNet show that ConvRoute-ViT improves computational efficiency while maintaining classification accuracy compared with the base ViT. 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/2610.38284 — 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 2026-11-01T02:16:43.875Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).