SWIFT: On-the-Fly Self-Speculative Decoding for LLM Inference Acceleration

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Xia, Heming, Li, Yongqi, Zhang, Jun, Du, Cunxiao, Li, Wenjie
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912261720768512
author Xia, Heming
Li, Yongqi
Zhang, Jun
Du, Cunxiao
Li, Wenjie
author_facet Xia, Heming
Li, Yongqi
Zhang, Jun
Du, Cunxiao
Li, Wenjie
contents Speculative decoding (SD) has emerged as a widely used paradigm to accelerate LLM inference without compromising quality. It works by first employing a compact model to draft multiple tokens efficiently and then using the target LLM to verify them in parallel. While this technique has achieved notable speedups, most existing approaches necessitate either additional parameters or extensive training to construct effective draft models, thereby restricting their applicability across different LLMs and tasks. To address this limitation, we explore a novel plug-and-play SD solution with layer-skipping, which skips intermediate layers of the target LLM as the compact draft model. Our analysis reveals that LLMs exhibit great potential for self-acceleration through layer sparsity and the task-specific nature of this sparsity. Building on these insights, we introduce SWIFT, an on-the-fly self-speculative decoding algorithm that adaptively selects intermediate layers of LLMs to skip during inference. SWIFT does not require auxiliary models or additional training, making it a plug-and-play solution for accelerating LLM inference across diverse input data streams. Our extensive experiments across a wide range of models and downstream tasks demonstrate that SWIFT can achieve over a 1.3x-1.6x speedup while preserving the original distribution of the generated text. We release our code in https://github.com/hemingkx/SWIFT.
format Preprint
id arxiv_https___arxiv_org_abs_2410_06916
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SWIFT: On-the-Fly Self-Speculative Decoding for LLM Inference Acceleration
Xia, Heming
Li, Yongqi
Zhang, Jun
Du, Cunxiao
Li, Wenjie
Computation and Language
Speculative decoding (SD) has emerged as a widely used paradigm to accelerate LLM inference without compromising quality. It works by first employing a compact model to draft multiple tokens efficiently and then using the target LLM to verify them in parallel. While this technique has achieved notable speedups, most existing approaches necessitate either additional parameters or extensive training to construct effective draft models, thereby restricting their applicability across different LLMs and tasks. To address this limitation, we explore a novel plug-and-play SD solution with layer-skipping, which skips intermediate layers of the target LLM as the compact draft model. Our analysis reveals that LLMs exhibit great potential for self-acceleration through layer sparsity and the task-specific nature of this sparsity. Building on these insights, we introduce SWIFT, an on-the-fly self-speculative decoding algorithm that adaptively selects intermediate layers of LLMs to skip during inference. SWIFT does not require auxiliary models or additional training, making it a plug-and-play solution for accelerating LLM inference across diverse input data streams. Our extensive experiments across a wide range of models and downstream tasks demonstrate that SWIFT can achieve over a 1.3x-1.6x speedup while preserving the original distribution of the generated text. We release our code in https://github.com/hemingkx/SWIFT.
title SWIFT: On-the-Fly Self-Speculative Decoding for LLM Inference Acceleration
topic Computation and Language
url https://arxiv.org/abs/2410.06916