Data processing method and device based on neural network

A neural network and data processing technology, applied in the field of data processing based on neural network, can solve the problem of reducing the demand for algorithm storage space, and achieve the effects of saving memory or hard disk space, reducing network bandwidth, and saving memory space

Inactive Publication Date: 2017-04-19
ALIBABA GRP HLDG LTD
View PDF0 Cites 65 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] One purpose of this application is to provide a data processing method and device based on neural network to solve the problem of reducing the storage space requirement of the algorithm without affecting the effect of the algorithm when implementing the neural network

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Data processing method and device based on neural network
  • Data processing method and device based on neural network
  • Data processing method and device based on neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0017] The application will be described in further detail below in conjunction with the accompanying drawings.

[0018] figure 1 A schematic flowchart showing a data processing method based on a neural network according to one aspect of the present application. like figure 1 As shown, the data processing method based on neural network includes:

[0019] Step S101, performing precision conversion on the single-precision floating-point data of the neural network.

[0020] Step S102, performing neural network calculation on the low-precision floating-point data formed through the precision conversion.

[0021] Wherein, the exponent item of the low-precision floating-point type data is smaller than the exponent item of the single-precision floating-point type data, and / or, the mantissa item of the low-precision floating-point type data is smaller than the single-precision floating-point type data mantissa item of .

[0022] Here, the neural network (NN, Neural Networks) in t...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The application aims to provide a data processing method and device based on a neural network. The method comprises steps that precision conversion for single precision floating point type data of the neural network is carried out; neural network calculation for the low precision floating point type data after precision conversion is carried out. Compared with the prior art, the method and the device are advantaged in that solving a large storage content problem in a low precision mode is facilitated, the memory space occupied by a model can be saved, a larger model can operate through utilizing same hardware configuration, and the memory or hardware space occupied by a data set can be saved; when the model is deployed on a cluster, network bandwidth required during synchronization can be effectively reduced, communication cost can be effectively reduced, and integral performance is improved.

Description

technical field [0001] The present application relates to the field of computers, in particular to a neural network-based data processing method and device. Background technique [0002] With the continuous upgrading of computer technology, neural networks have made great progress in simulating human intelligence. After decades of development, the current neural network is developing towards larger models and larger data sets. Using larger models and larger data sets can achieve higher classification and detection accuracy, but it brings problems. It is a substantial increase in the amount of calculation and storage. The problem of large amount of calculation can be solved by high-performance hardware such as multi-core CPU and GPU, using larger memory capacity, larger hard disk capacity, and faster network hardware to meet the increasing neural network parameters and data sets, but it is facing upgrades Long cycle, poor stability, excessive investment in equipment and oth...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06N3/02
CPCG06N3/02
Inventor 赵永科
Owner ALIBABA GRP HLDG LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap