Image de couverture de livre

DREAM: Drafting with Refined Target Features and Entropy-Adaptive Cross-Attention Fusion for Multimodal Speculative Decoding

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Hu, Yunhai, Xia, Tianhua, Liu, Zining, Raman, Rahul, Liu, Xingyu, Bao, Bo, Sather, Eric, Thangarasa, Vithursan, Zhang, Sai Qian
Format:	Preprint
Publié:	2025
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation
par: Liu, Zining, et autres
Publié: (2026)

DREAM-R: Multimodal Speculative Reasoning with RL-Based Refined Drafting, Precise Verification, and Fully Parallel Execution
par: Hu, Yunhai, et autres
Publié: (2026)

CodeQuant: Unified Clustering and Quantization for Enhanced Outlier Smoothing in Low-Precision Mixture-of-Experts
par: Yin, Xiangyang, et autres
Publié: (2026)

TapOut: A Bandit-Based Approach to Dynamic Speculative Decoding
par: Sridhar, Aditya, et autres
Publié: (2025)

MASSV: Multimodal Adaptation and Self-Data Distillation for Speculative Decoding of Vision-Language Models
par: Ganesan, Mugilan, et autres
Publié: (2025)

Speculative Decoding and Beyond: An In-Depth Survey of Techniques
par: Hu, Yunhai, et autres
Publié: (2025)

PEARL: Parallel Speculative Decoding with Adaptive Draft Length
par: Liu, Tianyu, et autres
Publié: (2024)

PipeSpec: Breaking Stage Dependencies in Hierarchical LLM Decoding
par: McDanel, Bradley, et autres
Publié: (2025)

OPT-Tree: Speculative Decoding with Adaptive Draft Tree Structure
par: Wang, Jikai, et autres
Publié: (2024)

PARD-2: Target-Aligned Parallel Draft Model for Dual-Mode Speculative Decoding
par: An, Zihao, et autres
Publié: (2026)

Learning to Draft: Adaptive Speculative Decoding with Reinforcement Learning
par: Zhang, Jiebin, et autres
Publié: (2026)

Sparse-IFT: Sparse Iso-FLOP Transformations for Maximizing Training Efficiency
par: Thangarasa, Vithursan, et autres
Publié: (2023)

Draft on the Fly: Adaptive Self-Speculative Decoding using Cosine Similarity
par: Metel, Michael R., et autres
Publié: (2024)

TAPS: Target-Aware Prefix Tree Selection for Diffusion-Drafted Speculative Decoding
par: Wang, Zhuoyu, et autres
Publié: (2026)

Self-Data Distillation for Recovering Quality in Pruned Large Language Models
par: Thangarasa, Vithursan, et autres
Publié: (2024)

TABED: Test-Time Adaptive Ensemble Drafting for Robust Speculative Decoding in LVLMs
par: Lee, Minjae, et autres
Publié: (2026)

FlexDraft: Flexible Speculative Decoding via Attention Tuning and Bonus-Guided Calibration
par: Zhang, Yaojie, et autres
Publié: (2026)

Draft Model Knows When to Stop: Self-Verification Speculative Decoding for Long-Form Generation
par: Zhang, Ziyin, et autres
Publié: (2024)

FastEagle: Cascaded Drafting for Accelerating Speculative Decoding
par: Huang, Haiduo, et autres
Publié: (2025)

Cost-Aware Diffusion Draft Trees for Speculative Decoding
par: Zhang, Shuai, et autres
Publié: (2026)

Accelerating Speculative Decoding with Block Diffusion Draft Trees
par: Ringel, Liran, et autres
Publié: (2026)

AHASD: Asynchronous Heterogeneous Architecture for LLM Adaptive Drafting Speculative Decoding on Mobile Devices
par: Zirui, Ma, et autres
Publié: (2026)

SJD-PAC: Accelerating Speculative Jacobi Decoding via Proactive Drafting and Adaptive Continuation
par: Kang, Jialiang, et autres
Publié: (2026)

AdaEAGLE: Optimizing Speculative Decoding via Explicit Modeling of Adaptive Draft Structures
par: Zhang, Situo, et autres
Publié: (2024)

OmniDraft: A Cross-vocabulary, Online Adaptive Drafter for On-device Speculative Decoding
par: Ramakrishnan, Ramchalam Kinattinkara, et autres
Publié: (2025)

Ouroboros: Generating Longer Drafts Phrase by Phrase for Faster Speculative Decoding
par: Zhao, Weilin, et autres
Publié: (2024)

Draft Less, Retrieve More: Hybrid Tree Construction for Speculative Decoding
par: Shen, Yuhao, et autres
Publié: (2026)

SpecBlock: Block-Iterative Speculative Decoding with Dynamic Tree Drafting
par: Shi, Weijie, et autres
Publié: (2026)

DFlare: Scaling Up Draft Capacity for Block Diffusion Speculative Decoding
par: Zhang, Jiebin, et autres
Publié: (2026)

TETRIS: Optimal Draft Token Selection for Batch Speculative Decoding
par: Wu, Zhaoxuan, et autres
Publié: (2025)

MineDraft: A Framework for Batch Parallel Speculative Decoding
par: Tang, Zhenwei, et autres
Publié: (2026)

When Drafts Evolve: Speculative Decoding Meets Online Learning
par: Qian, Yu-Yang, et autres
Publié: (2026)

SpecHub: Provable Acceleration to Multi-Draft Speculative Decoding
par: Sun, Ryan, et autres
Publié: (2024)

Draft, Verify, and Improve: Toward Training-Aware Speculative Decoding
par: Bhansali, Shrenik, et autres
Publié: (2025)

POSS: Position Specialist Generates Better Draft for Speculative Decoding
par: Huang, Langlin, et autres
Publié: (2025)

REAP the Experts: Why Pruning Prevails for One-Shot MoE compression
par: Lasby, Mike, et autres
Publié: (2025)

SD$^2$: Self-Distilled Sparse Drafters
par: Lasby, Mike, et autres
Publié: (2025)

MediSwift: Efficient Sparse Pre-trained Biomedical Language Models
par: Thangarasa, Vithursan, et autres
Publié: (2024)

FlexSpec: Frozen Drafts Meet Evolving Targets in Edge-Cloud Collaborative LLM Speculative Decoding
par: Li, Yuchen, et autres
Publié: (2026)

SpecBranch: Speculative Decoding via Hybrid Drafting and Rollback-Aware Branch Parallelism
par: Shen, Yuhao, et autres
Publié: (2025)