Falcon: Faster and Parallel Inference of Large Language Models Through Enhanced Semi-Autoregressive Drafting and Custom-Designed Decoding Tree

Open Webpage

Xiangxiang Gao, Weisheng Xie, Yiwei Xiang, Feng Ji

Published: 01 Jan 2025, Last Modified: 15 May 2025AAAI 2025EveryoneRevisionsBibTeXCC BY-SA 4.0