Method and device for optimizing image neural network structure
A technology of network structure and optimization method, applied in the optimization field of image neural network structure, can solve the problems of low degree of automation, high use threshold, unfavorable image neural network structure, etc.
Pending Publication Date: 2021-05-18
SHANGHAI YITU NETWORK SCI & TECH
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] The embodiment of the present application provides a method and device for optimizing an image neural network structure, which is used to solve the problem that the neural network structure search scheme in the prior art has a relatively low degree of automation and a relatively high threshold for use, which is not conducive to quickly obtaining an image neural network structure. question
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0068]In order to solve the problem of automation of neural network structure search scheme in the prior art, the use of threshold is relatively high, which is unfavorable to quickly obtain the image neural network structure, the present application provides an optimization method of image neural network structure and Device.
[0069]The preferred embodiment of the present application will be described with reference to the accompanying drawings, and the preferred embodiments described herein are intended to illustrate and explain the present application, and is not intended to limit the present application, and in the case of unlicy, the present application The features in the embodiments and examples in the examples can be combined with each other.
[0070]Neural Architecture Search (NAS) refers to the use of a circular neural network as a controller to generate a subnecical neural network as a controller in the case of uncertain neural network, and then conduct training and evaluation ...
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
Login to View More Abstract
The invention discloses a method and device for optimizing an image neural network structure, and belongs to the technical field of artificial intelligence. The method comprises the steps: generating a first image neural network structure according to structure description information contained in a received image neural network structure optimization request, comparing the first image neural network structure with each neural network structure template so as to determine a target neural network structure template used by the first image neural network structure and a target sub-neural network structure corresponding to the target neural network structure template in the first image neural network structure; using a plurality of preset image neural network structures corresponding to each target neural network structure template to replace a target sub-neural network structure corresponding to the target neural network structure template in the first image neural network structure to obtain a candidate neural network structure set; and performing neural network structure search in the candidate neural network structure set to obtain an optimized second image neural network structure.
Description
Technical field[0001]The present application relates to artificial intelligence technology, in particular, to an optimization method and apparatus for image neural network structure.Background technique[0002]In the field of artificial intelligence technology, deep learning technology is an important branch, and during the development of deep learning technology, image neural network structural design has always played a crucial role.[0003]In recent years, the image neural network structure designed by the field experts has achieved huge success in various tasks. However, as the actual business needs are increasing, the existing image neural network structure is increasingly difficult to meet the actual business needs, so the purpose of neural architecture search, NAS, neural network structure is automated. Design image neural network structure, saving human cost and time cost. However, both existing neural network structural search schemes require users to have some understanding of...
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
Login to View More IPC IPC(8): G06N3/04
CPCG06N3/04
Inventor 任广辉谢文韬陈云鹏
Owner SHANGHAI YITU NETWORK SCI & TECH