公开(公告)号：US20250139422A1

公开(公告)日：2025-05-01

申请号：US18919294

申请日：2024-10-17

Applicant: Samsung Electronics Co., Ltd.

Inventor： Shanshan LV ,  Jonghoon YOON ,  Byung In YOO ,  Changyong SON ,  Sung-Jae CHO ,  Yunhao ZHANG ,  Zhenxin YANG ,  Miao ZHANG

IPC: G06N3/0495

Abstract: A method performed by one or more processors includes: iteratively training layer-specific quantization levels and layer-specific quantization intervals of respective layers of a neural network of original weights by, for each training iteration, adjusting the quantization levels and quantization intervals to reduce a loss that is determined based on the original weights and is determined based on the original weights as quantized according to the quantization levels and quantization intervals at a current iteration of the training.