Unlock instant, AI-driven research and patent intelligence for your innovation.

Method and device for realizing deep convolutional neural network processing

A deep convolution, neural network technology, applied in the computer field, can solve problems such as algorithms that do not consider acceleration problems

Pending Publication Date: 2021-02-26
ZTE CORP
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the deep convolutional neural network after pruning and quantization still has the potential to accelerate. The current algorithm does not consider the acceleration after pruning and quantization, so how to accelerate the network after pruning and quantization remains to be studied.

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
  • Method and device for realizing deep convolutional neural network processing
  • Method and device for realizing deep convolutional neural network processing
  • Method and device for realizing deep convolutional neural network processing

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0057] In a typical configuration of the present application, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0058] Memory may include non-permanent storage in computer-readable media, in the form of random access memory (RAM) and / or nonvolatile memory, such as read-only memory (ROM) or flash memory (flashRAM). Memory is an example of computer readable media.

[0059] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can be implemented by any method or technology for storage of information. Information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Progr...

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 invention discloses a method and device for realizing deep convolutional neural network processing, and the method comprises the steps of carrying out the partitioning of quantized parameters, anddividing each convolution kernel of each convolution layer of a deep convolutional neural network into a plurality of sub-convolution kernels; and counting the category number of all parameters in each sub convolution kernel. According to the invention, a guarantee is provided for better realizing acceleration of the deep convolutional neural network. Furthermore, for each sub convolution kernel,clustering processing is carried out according to the number of parameters in each category. In this way, acceleration of the deep convolutional neural network is better realized.

Description

technical field [0001] This application relates to but is not limited to the field of computer technology, especially a method and device for implementing deep convolutional neural network processing. Background technique [0002] Deep convolutional neural network (CNN, Convolutional Neural Network) has achieved remarkable results in the fields of computer vision, speech recognition, natural language processing and bioinformatics, and has significantly improved image classification, target detection and tracking, semantic segmentation, The performance of tasks such as face recognition and video analysis. CNN has also been widely used in industries such as autonomous driving, smart medical care, security, and the Internet. [0003] Most deep convolutional neural networks have millions of parameters and calculations. For example, the AlexNet network has 60M parameters and 720M multiply-add operations; another example: the VGG16 network has 138M parameters and 15300M multiply-...

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): G06N3/04G06N3/08G06K9/62
CPCG06N3/08G06N3/045G06F18/23
Inventor 杨德友刘文鸿张欣叶蕾
Owner ZTE CORP