Metraipolis › archive › dc_0c5dd62ff342

Llama-Mobile: Efficient 2.7-Bit Quantization of VLMs — arXiv:2608.21134

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.21134
License (as archived)
CC0-1.0
Archived
2026-10-18T10:31:31.649Z
SHA-256 of the archived text
22639d25723a1c4afac11eec67515d15bf547942decbf4d11aa09a07dbc9d254
Stable URL of this page
https://metraipolis.com/documento/dc_0c5dd62ff342

How to cite

Llama-Mobile: Efficient 2.7-Bit Quantization of VLMs — arXiv:2608.21134 — arXiv (en), https://arxiv.org/abs/2608.21134 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0c5dd62ff342 (Metraipolis archive dc_0c5dd62ff342, sha256 22639d25723a1c4afac11eec67515d15bf547942decbf4d11aa09a07dbc9d254)
BibTeX
@misc{dc0c5dd62ff342,
  title = {Llama-Mobile: Efficient 2.7-Bit Quantization of VLMs — arXiv:2608.21134},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.21134},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0c5dd62ff342; sha256 22639d25723a1c4afac11eec67515d15bf547942decbf4d11aa09a07dbc9d254},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.21134. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Deploying vision-language models (VLMs) on mobile devices is challenging due to their significant memory and compute requirements. We present a framework for quantizing VLMs for efficient inference on resource-constrained hardware. Our approach combines a quantization pipeline that uses the model itself to generate training data and does not require access to the training setup, with a novel 2.7-bit-per-parameter format supporting efficient execution on Arm CPUs. We validate our approach by compressing the Llama 3.2 11B Vision Instruct model to 3.7 GB with 8-bit activations, preserving strong performance on a set of standard visual question answering tasks.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-24 in cs.LG).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.21134 — 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 2026-10-18T10:27:56.002Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).