DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Liu, Zining, Hu, Yunhai, Xia, Tianhua, Bao, Bo, Sather, Eric, Thangarasa, Vithursan, Zhang, Sai Qian
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918533537988608
author Liu, Zining
Hu, Yunhai
Xia, Tianhua
Bao, Bo
Sather, Eric
Thangarasa, Vithursan
Zhang, Sai Qian
author_facet Liu, Zining
Hu, Yunhai
Xia, Tianhua
Bao, Bo
Sather, Eric
Thangarasa, Vithursan
Zhang, Sai Qian
contents Speculative decoding (SD) has proven to be an effective technique for accelerating autoregressive generation in large language models (LLMs) however, its application to vision-language models (VLMs) remains relatively unexplored. We propose~\textit{DREAM-S}, a novel SD framework designed specifically for fast and efficient decoding in VLMs. DREAM-S leverages a neural architecture search (NAS) framework with target-aware supernet training to automatically identify both the optimal interaction strategy between the draft and target models, and the most suitable draft model architecture for the underlying hardware implementation platform. DREAM-S additionally incorporates adaptive intermediate feature distillation, guided by attention entropy, to enable efficient draft training. Experiments on a range of well-established VLMs show that DREAM-S achieves up to a $3.85\times$ speedup compared to standard decoding approaches and significantly outperforms existing SD baselines. The code is publicly available at: https://github.com/SAI-Lab-NYU/DREAM-S .
format Preprint
id arxiv_https___arxiv_org_abs_2606_00535
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation
Liu, Zining
Hu, Yunhai
Xia, Tianhua
Bao, Bo
Sather, Eric
Thangarasa, Vithursan
Zhang, Sai Qian
Machine Learning
Speculative decoding (SD) has proven to be an effective technique for accelerating autoregressive generation in large language models (LLMs) however, its application to vision-language models (VLMs) remains relatively unexplored. We propose~\textit{DREAM-S}, a novel SD framework designed specifically for fast and efficient decoding in VLMs. DREAM-S leverages a neural architecture search (NAS) framework with target-aware supernet training to automatically identify both the optimal interaction strategy between the draft and target models, and the most suitable draft model architecture for the underlying hardware implementation platform. DREAM-S additionally incorporates adaptive intermediate feature distillation, guided by attention entropy, to enable efficient draft training. Experiments on a range of well-established VLMs show that DREAM-S achieves up to a $3.85\times$ speedup compared to standard decoding approaches and significantly outperforms existing SD baselines. The code is publicly available at: https://github.com/SAI-Lab-NYU/DREAM-S .
title DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation
topic Machine Learning
url https://arxiv.org/abs/2606.00535