Speculate Deep and Accurate: Lossless and Training-Free Acceleration for Offloaded LLMs via Substitute Speculative Decoding

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Wang, Pei-Shuo, Chen, Jian-Jia, Yang, Chun-Che, Chang, Chi-Chih, Huang, Ning-Chi, Abdelfattah, Mohamed S., Wu, Kai-Chiang
Format:	Preprint
Published:	2025
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2509.18344