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

GAN network compression method, apparatus and device, and storage medium

A compression method and network technology, applied in neural learning methods, biological neural network models, neural architectures, etc.

Pending Publication Date: 2021-03-09
IFLYTEK CO LTD
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, the current image generation network is several orders of magnitude more computationally intensive than the recognition 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
  • GAN network compression method, apparatus and device, and storage medium
  • GAN network compression method, apparatus and device, and storage medium
  • GAN network compression method, apparatus and device, and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the application with reference to the drawings in the embodiments of the application. Apparently, the described embodiments are only some of the embodiments of the application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

[0059] In order to deploy GAN networks to lower-performance end devices, GAN networks need to be compressed. The applicant in this case first tried to compress the generation network G in the GAN network separately, keeping the discriminant network D unchanged, but found that this compression method would destroy the balance of the GAN network, and it was difficult to obtain the desired effect. Analysis of the reasons shows that , the GAN network is trained against the generation ne...

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 GAN network compression method, apparatus and device, and a storage medium. The method comprises the steps of: for a to-be-compressed GAN network, repalcing an original operator of each layer with at least one candidate operator, wherein the output of each layer is composed of the output of each candidate operator after replacement of each layer, and the calculated amountof the candidate operators is less than calculated amount of replaced original operator; initializing parameters and weights of the candidate operators in the GAN network, alternately updating a generation network G and a discrimination network D until the GAN network converges, and on the basis of the final weights of the candidate operators in each layer in the convergent GAN network, determining the finally reserved operator of each layer to obtain a compressed GAN network. According to the scheme of the invention, the generation network G and the discrimination network D in the GAN network are compressed at the same time, and the balance of the compressed GAN network is ensured while the GAN network is compressed.

Description

technical field [0001] The present application relates to the technical field of model processing, and more specifically, relates to a GAN network compression method, device, equipment and storage medium. Background technique [0002] With the development of image and video generation technology, GAN-based generation technology has been widely used, such as image conversion, image restoration, image super-resolution, etc. [0003] GAN (Generative Adversarial Networks, Generative Adversarial Networks) is a deep learning model and one of the most promising methods for unsupervised learning on complex distributions in recent years. The GAN network produces a fairly good output through the mutual game learning of the generation network G (Generative Model) and the discriminant network D (Discriminative Model). [0004] But current image generation networks are orders of magnitude more computationally intensive than recognition networks. In order to deploy the generated network...

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): G06N3/04G06N3/08
CPCG06N3/082G06N3/045
Inventor 左童春何山胡金水刘聪殷兵
Owner IFLYTEK CO LTD
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com