Metraipolis › archive › dc_61bcdb5f68dc

PACE: A Unified Condense-and-Extract Paradigm for Fast VLM Inference — arXiv:2608.27206

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.27206
License (as archived)
CC0-1.0
Archived
2026-10-17T14:30:41.797Z
SHA-256 of the archived text
b63b6ff2571042776544b3576e667ef70a67d8a69853d263d0263940f82be8b3
Stable URL of this page
https://metraipolis.com/documento/dc_61bcdb5f68dc

How to cite

PACE: A Unified Condense-and-Extract Paradigm for Fast VLM Inference — arXiv:2608.27206 — arXiv (en), https://arxiv.org/abs/2608.27206 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_61bcdb5f68dc (Metraipolis archive dc_61bcdb5f68dc, sha256 b63b6ff2571042776544b3576e667ef70a67d8a69853d263d0263940f82be8b3)
BibTeX
@misc{dc61bcdb5f68dc,
  title = {PACE: A Unified Condense-and-Extract Paradigm for Fast VLM Inference — arXiv:2608.27206},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.27206},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_61bcdb5f68dc; sha256 b63b6ff2571042776544b3576e667ef70a67d8a69853d263d0263940f82be8b3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.27206. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-Language Models (VLMs) demonstrate exceptional visual reasoning capabilities, yet their inference costs escalate rapidly with the proliferation of visual tokens. Existing visual token pruning methods exhibit two fundamental limitations. First, most approaches operate exclusively post-vision encoder, leaving the substantial latency of the visual encoding phase unoptimized. Second, under strict token budgets, these methods often fail to jointly preserve holistic visual contexts and fine-grained details, leading to performance degradation. To address these bottlenecks, we propose PACE (Pixel-Adaptive Condense and Extract), a training-free inference framework that accelerates both the vision encoder and the Large Language Model (LLM) via a unified Condense-and-Extract paradigm. During the Condense stage, an Adaptive Pixel Compressor (APC) evaluates visual information density prior to encoding, adaptively downsampling redundant inputs, curtailing encoder computation while preserving global context and essential visual cues. In the Extract stage, a Dynamic Dual-Attention Extractor (DDAE) selectively retains visual tokens via a fusion of internal visual signals from the encoder and semantic signals from the LLM, safeguarding task-critical details. By integrating PACE into Qwen2.5-VL-7B, the model retains 93.8% of its original performance while utilizing only 10% of the visual tokens, yielding a 3.1x speedup in time to first token (TTFT). Our code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-28 in cs.AI).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.27206 — 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 2026-10-17T13:16:53.699Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).