Metraipolis › archive › dc_2d5c5671e1df
PETR: Prompt Ensembling with Training-free Routing for Vision-Language Models — arXiv:2609.23600
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.23600
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T21:59:49.216Z
- SHA-256 of the archived text
ad6324f3b68a5ed887476851ca4e259dbd427a6695ef2bc4f6ba799b15a6e40d- Stable URL of this page
https://metraipolis.com/documento/dc_2d5c5671e1df
How to cite
PETR: Prompt Ensembling with Training-free Routing for Vision-Language Models — arXiv:2609.23600 — arXiv (en), https://arxiv.org/abs/2609.23600 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2d5c5671e1df (Metraipolis archive dc_2d5c5671e1df, sha256 ad6324f3b68a5ed887476851ca4e259dbd427a6695ef2bc4f6ba799b15a6e40d)
BibTeX
@misc{dc2d5c5671e1df,
title = {PETR: Prompt Ensembling with Training-free Routing for Vision-Language Models — arXiv:2609.23600},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.23600},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2d5c5671e1df; sha256 ad6324f3b68a5ed887476851ca4e259dbd427a6695ef2bc4f6ba799b15a6e40d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.23600. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Prompt learning efficiently adapts vision-language models (VLMs) to downstream tasks, but gains on seen classes often come at the expense of generalization to unseen classes. To address this limitation, we propose prompt ensembling with training-free routing (PETR), whose key innovation is a carefully designed dual-prompt architecture: two complementary prompts are learned from different data and objectives to emphasize seen class discrimination and unseen-class generalization, respectively. During training, both prompts are fine-tuned using a shared frozen CLIP backbone, and statistical information is collected from the training set logits. At inference time, we determine the similarity of each test sample to seen data, and route the sample to the most appropriate prompt branch. To the best of our knowledge, this is the first prompt tuning framework that performs training-free adaptive routing based on statistical similarity. This design provides an interpretable routing signal and avoids common MoE-style routing pathologies, such as router training instability and load imbalance. Extensive experiments on 11 benchmark datasets demonstrate that our framework consistently outperforms previous methods on both seen and unseen classes, achieving new state-of-the-art results. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.23600 — 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 2026-10-09T21:45:20.932Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).