Processing

Please wait...

Settings

Settings

Goto Application

1. WO2021057746 - NEURAL NETWORK PROCESSING METHOD AND APPARATUS, COMPUTER DEVICE AND STORAGE MEDIUM

Publication Number WO/2021/057746
Publication Date 01.04.2021
International Application No. PCT/CN2020/116933
International Filing Date 22.09.2020
IPC
G06N 3/063 2006.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
06Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
063using electronic means
CPC
G06N 3/063
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
06Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
063using electronic means
Applicants
  • 安徽寒武纪信息科技有限公司 ANHUI CAMBRICON INFORMATION TECHNOLOGY CO., LTD. [CN]/[CN]
Inventors
  • 张潇 ZHANG, Xiao
  • 周玉松 ZHOU, Yusong
  • 孟小甫 MENG, Xiaofu
Agents
  • 广州三环专利商标代理有限公司 SCIHEAD IP LAW FIRM
Priority Data
201910910117.624.09.2019CN
201910910118.024.09.2019CN
Publication Language Chinese (ZH)
Filing Language Chinese (ZH)
Designated States
Title
(EN) NEURAL NETWORK PROCESSING METHOD AND APPARATUS, COMPUTER DEVICE AND STORAGE MEDIUM
(FR) PROCÉDÉ ET APPAREIL DE TRAITEMENT DE RÉSEAU NEURONAL, DISPOSITIF INFORMATIQUE ET SUPPORT DE STOCKAGE
(ZH) 神经网络处理方法、装置、计算机设备及存储介质
Abstract
(EN)
Disclosed are a neural network processing method and apparatus, a computer device and a storage medium. An operator is split into a plurality of smaller-scale operators, such that a multi-core processor can directly call a calculation library under a single-core architecture, whereby a hardware resource of the multi-core processor is used fully.
(FR)
L'invention concerne un procédé et un appareil de traitement de réseau neuronal, un dispositif informatique et un support de stockage. Un opérateur est divisé en une pluralité d'opérateurs plus petits, de telle sorte qu'un processeur multicœur peut appeler directement une bibliothèque de calcul dans une architecture à noyau unique, moyennant quoi une ressource matérielle du processeur multicœur est utilisée complètement.
(ZH)
本申请实施例公开了一种神经网络处理方法、装置、计算机设备及存储介质,通过将算子拆分成多个规模更小的算子,这样多核处理器可以直接调用单核架构下的计算库,充分利用了多核处理器的硬件资源。
Latest bibliographic data on file with the International Bureau