Convolutional neural network processing method and device and electronic system

A convolutional neural network and processing method technology, applied in the field of convolutional neural network processing methods, devices and electronic systems, can solve the problem of degradation, poor performance of lightweight convolutional neural networks, and lightweight convolutional neural network expression ability Insufficient and other problems, to achieve the effect of improving network accuracy, reducing training time, and improving expression ability

Pending Publication Date: 2020-04-28
MEGVII BEIJINGTECH CO LTD
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, since the amount of parameters and calculations of the lightweight convolutional neural network are greatly reduced compared with the ordinary convolutional neural network, th

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
  • Convolutional neural network processing method and device and electronic system
  • Convolutional neural network processing method and device and electronic system
  • Convolutional neural network processing method and device and electronic system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] First, refer to figure 1An example electronic system 100 for implementing the convolutional neural network processing method, device and electronic system of the embodiments of the present invention will be described.

[0034] Such as figure 1 A schematic structural diagram of an electronic system is shown, the electronic system 100 includes one or more processing devices 102, one or more storage devices 104, input devices 106, output devices 108 and one or more image acquisition devices 110, these components The interconnections are via bus system 112 and / or other forms of connection mechanisms (not shown). It should be noted that figure 1 The components and structures of the electronic system 100 shown are exemplary rather than limiting, and the electronic system may also have other components and structures as required.

[0035] The processing device 102 may be a smart terminal, or a device including a central processing unit (CPU) or other forms of processing uni...

Embodiment 2

[0042] This embodiment provides a convolutional neural network processing method, which improves the traditional convolutional neural network. In this embodiment, the second convolutional neural network is the target in the first convolutional neural network. The convolution kernel is extended to the network after the first convolution kernel and the second convolution kernel, where the number of input channels of the first convolution kernel is the same as the number of input channels of the target convolution kernel, and the output channel of the second convolution kernel The number is the same as the number of output channels of the target convolution kernel.

[0043] see figure 2 A schematic diagram of a typical convolutional structure of the first convolutional neural network and image 3 A schematic diagram of a typical convolutional structure of the second convolutional neural network shown, such as figure 2 As shown, for an input feature layer, the target convoluti...

Embodiment 3

[0061] This embodiment provides another convolutional neural network processing method, which is implemented on the basis of the above-mentioned embodiments; this embodiment focuses on initializing the second convolutional neural network based on the parameters of the trained first convolutional neural network parameters of the step. Such as Figure 5 The flow chart of another convolutional neural network processing method is shown. The convolutional neural network processing method in this embodiment includes the following steps:

[0062] Step S502, train the first convolutional neural network, and obtain the parameters of the trained first convolutional neural network; the above parameters include weight parameters and first-type parameters; the first-type parameters include partial values ​​of the first convolutional neural network. parameters and parameters in the batch normalization layer of the first convolutional neural network.

[0063] Classify the parameters of the...

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 provides a convolutional neural network processing method and device and an electronic system. The method comprises the following steps: training a first convolutional neural network, and initializing parameters of a second convolutional neural network based on the trained parameters of the first convolutional neural network; training the second convolutional neural network; and performing structure conversion on the trained second convolutional neural network to enable the structure of the converted second convolutional neural network to be the same as the structure of the firstconvolutional neural network. In the mode, the convolution kernel of the first convolution neural network is expanded to obtain the second convolution neural network, so that the expression capability of the network in the network training process can be improved; the parameters of the second convolutional neural network are initialized based on the trained parameters of the first convolutional neural network, so that the training time of the network can be reduced; and the structure of the second convolutional neural network is converted into the same structure of the first convolutional neural network, so that the network operation speed is not influenced while the network precision is improved.

Description

technical field [0001] The present invention relates to the technical field of convolutional neural networks, in particular to a convolutional neural network processing method, device and electronic system. Background technique [0002] Convolutional neural networks are being developed and applied rapidly. In order to pursue high performance, the depth and complexity of convolutional neural network models are increasing. However, in real application scenarios such as mobile devices or embedded devices, such large and complex Convolutional neural models are difficult to apply. [0003] In related technologies, limited by insufficient memory and real-time requirements, only lightweight convolutional neural networks can be used in actual scenarios. However, since the amount of parameters and calculations of the lightweight convolutional neural network are greatly reduced compared with the ordinary convolutional neural network, the expressive ability of the lightweight convolut...

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
IPC IPC(8): G06N3/04
CPCG06N3/045
Inventor 李志远李伯勋俞刚
Owner MEGVII BEIJINGTECH CO 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