DuoQ: A DSP Utilization-aware and Outlier-free Quantization for FPGA-based LLMs Acceleration.

Saved in:
Bibliographic Details
Title: DuoQ: A DSP Utilization-aware and Outlier-free Quantization for FPGA-based LLMs Acceleration.
Authors: Yu, Zhuoquan1, zqyu21@m.fudan.edu.cn, Ji, Huidong1, hdji21@m.fudan.edu.cn, Cao, Yue1, caoyue23@m.fudan.edu.cn, Wu, Junfu1, junfuwu23@m.fudan.edu.cn, Yan, Xiaoze1, xzyan23@m.fudan.edu.cn, Zheng, Lirong1, lrzheng@fudan.edu.cn, Zou, Zhuo1, zhuo@fudan.edu.cn
Source: DAC: Annual ACM/IEEE Design Automation Conference; 2025, Issue 62, p1-7, 7p
Database: Applied Science & Technology Source
Description
ISSN:0738100X
DOI:10.1109/DAC63849.2025.11132816