English
Related papers

Related papers: PaSS: Parallel Speculative Sampling

200 papers

Diffusion Large Language Models (dLLMs) offer fast, parallel token generation, but their standalone use is plagued by an inherent efficiency-quality tradeoff. We show that, if carefully applied, the attributes of dLLMs can actually be a…

Machine Learning · Computer Science 2026-01-29 Rui Pan , Zhuofu Chen , Hongyi Liu , Arvind Krishnamurthy , Ravi Netravali

Autoregressive language models demonstrate excellent performance in various scenarios. However, the inference efficiency is limited by its one-step-one-word generation mode, which has become a pressing problem recently as the models become…

Computation and Language · Computer Science 2025-04-25 Jikai Wang , Yi Su , Juntao Li , Qingrong Xia , Zi Ye , Xinyu Duan , Zhefeng Wang , Min Zhang

Tensor parallelism is an essential technique for distributed training of large neural networks. However, automatically determining an optimal tensor parallel strategy is challenging due to the gigantic search space, which grows…

Machine Learning · Computer Science 2025-08-06 Ziji Shi , Le Jiang , Ang Wang , Jie Zhang , Chencan Wu , Yong Li , Xiaokui Xiao , Wei Lin , Jialin Li

Speculative decoding has been shown as an effective way to accelerate Large Language Model (LLM) inference by using a Small Speculative Model (SSM) to generate candidate tokens in a so-called speculation phase, which are subsequently…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-03-21 Fahao Chen , Peng Li , Tom H. Luan , Zhou Su , Jing Deng

We propose an acceleration scheme for large language models (LLMs) through Speculative Decoding with Semantic Adaptive Tokens (SDSAT). The primary objective of this design is to enhance the LLM model's ability to generate draft tokens more…

Computation and Language · Computer Science 2024-04-02 Chengbo Liu , Yong Zhu

Large language models increasingly rely on explicit reasoning chains and can produce multiple plausible responses for a given context. We study the candidate sampler that produces the set of plausible responses contrasting the ancestral…

Computation and Language · Computer Science 2025-09-23 Sergey Troshin , Irina Saparina , Antske Fokkens , Vlad Niculae

The dominant approach to generating from language models subject to some constraint is locally constrained decoding (LCD), incrementally sampling tokens at each time step such that the constraint is never violated. Typically, this is…

Token-based masked generative models are gaining popularity for their fast inference time with parallel decoding. While recent token-based approaches achieve competitive performance to diffusion-based models, their generation performance is…

Machine Learning · Computer Science 2023-04-05 Jaewoong Lee , Sangwon Jang , Jaehyeong Jo , Jaehong Yoon , Yunji Kim , Jin-Hwa Kim , Jung-Woo Ha , Sung Ju Hwang

Autoregressive (AR) language models generate text one token at a time, even when consecutive tokens are highly predictable given earlier context. We introduce MARS (Mask AutoRegreSsion), a lightweight fine-tuning method that teaches an…

Computation and Language · Computer Science 2026-04-09 Ziqi Jin , Lei Wang , Ziwei Luo , Aixin Sun

We present a novel inference scheme, self-speculative decoding, for accelerating Large Language Models (LLMs) without the need for an auxiliary model. This approach is characterized by a two-stage process: drafting and verification. The…

Computation and Language · Computer Science 2025-02-11 Jun Zhang , Jue Wang , Huan Li , Lidan Shou , Ke Chen , Gang Chen , Sharad Mehrotra

Speculative Decoding is a prominent technique for accelerating the autoregressive inference of large language models (LLMs) by employing a fast draft model to propose candidate token sequences and a large target model to verify them in…

Computation and Language · Computer Science 2025-12-18 Chendong Sun , Ali Mao , Lei Xu , mingmin Chen

Scaling test-time compute brings substantial performance gains for large language models (LLMs). By sampling multiple answers and heuristically aggregate their answers (e.g., either through majority voting or using verifiers to rank the…

Computation and Language · Computer Science 2025-10-13 Jianing Qi , Xi Ye , Hao Tang , Zhigang Zhu , Eunsol Choi

Large language models and large multimodal models (LLMs and LMMs) deliver strong generative performance but suffer from slow decoding, a problem that becomes more severe when handling visual inputs, whose sequences typically contain many…

Computer Vision and Pattern Recognition · Computer Science 2026-02-04 Zihua Wang , Ruibo Li , Haozhe Du , Joey Tianyi Zhou , Yu Zhang , Xu Yang

The recent advancements in large language models (LLMs) have been extraordinary, yet the escalating inference costs associated with them present challenges in real-world applications. To address these challenges, we propose a novel approach…

Computation and Language · Computer Science 2024-06-07 Jiahao Liu , Qifan Wang , Jingang Wang , Xunliang Cai

Many applications today provide users with multiple auto-complete drafts as they type, including GitHub's code completion, Gmail's smart compose, and Apple's messaging auto-suggestions. Under the hood, language models support this by…

Speculative decoding accelerates LLM inference by using a fast draft model to generate tokens and a more accurate target model to verify them. Its performance depends on the $\textit{acceptance length}$, or number of draft tokens accepted…

Computation and Language · Computer Science 2026-05-20 Avinash Kumar , Sujay Sanghavi , Poulami Das

We propose TETRIS, a novel method that optimizes the total throughput of batch speculative decoding in multi-request settings. Unlike existing methods that optimize for a single request or a group of requests as a whole, TETRIS actively…

Computation and Language · Computer Science 2025-06-02 Zhaoxuan Wu , Zijian Zhou , Arun Verma , Alok Prakash , Daniela Rus , Bryan Kian Hsiang Low

Speculative decoding has become the standard approach for accelerating Large Language Model (LLM) inference. It exploits a lossless draft-then-verify procedure to circumvent the latency of autoregressive decoding, achieving impressive…

Computation and Language · Computer Science 2025-11-05 Jameson Sandler , Jacob K. Christopher , Thomas Hartvigsen , Ferdinando Fioretto

In autoregressive language models, each token is sampled by conditioning on all the past tokens; the overall string has thus been sampled from the correct underlying joint distribution represented by the model. In contrast, masked diffusion…

Computation and Language · Computer Science 2026-02-03 Parikshit Bansal , Sujay Sanghavi

Speculative decoding accelerates large language model inference by proposing tokens with a lightweight draft model and selectively accepting them using a target model. This work introduces DropMatch, a novel approach that matches draft…

Computation and Language · Computer Science 2026-03-05 Jeongtae Lee , Minjung Jo , Hyunjoon Jeong , Gunho Park , Sunghyeon Woo , Joonghoon Kim , Se Jung Kwon , Dongsoo Lee