DySpec: Faster speculative decoding with dynamic token tree structure.

Yunfan Xiong|Ruoyu Zhang|Yanzeng Li|Lei Zou


Anthology ID:DBLP:journals/www/XiongZLZ25
Year:2025
Venue:World Wide Web (WWW)
Pages:36
URL:https://doi.org/10.1007/S11280-025-01344-0
DOI:https://doi.org/10.1007/S11280-025-01344-0
DBLP:journals/www/XiongZLZ25
BibTeX:
@article{xiong-2025-dyspec, author = {Yunfan Xiong and Ruoyu Zhang and Yanzeng Li and Lei Zou}, title = {{DySpec: Faster speculative decoding with dynamic token tree structure}}, journal = {World Wide Web (WWW)}, volume = {28}, number = {3}, pages = {36}, year = {2025}, url = {https://doi.org/10.1007/S11280-025-01344-0}, doi = {https://doi.org/10.1007/S11280-025-01344-0} }