TMA-Adaptive FP8 Grouped GEMM: Eliminating Padding Requirements in Low-Precision Training and Inference on Hopper

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Su, Zhongling, Fu, Rong, Cao, Weihan, Gao, Jianfei, Jin, Minxi, Pei, Zhilin, Wang, Hui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!