Deterministic Inference across Tensor Parallel Sizes That Eliminates Training-Inference Mismatch

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Ziyang, Ding, Xinheng, Yuan, Jiayi, Liu, Rixin, Mao, Huizi, Xing, Jiarong, Liu, Zirui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!