Deep learning model deployment method and device, and medium

A deep learning and model technology, applied in the field of deep learning model deployment, can solve the problems of data sharing, difficult deployment, and difficult landing, and achieve the effects of fast inference, good stability, and good versatility

Inactive Publication Date: 2022-06-24
苏州中科行智智能科技有限公司
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] In recent years, the advantages of deep learning in the direction of computer vision have become more and more obvious, and industrial visual inspection based on deep learning has also received more and more attention. However, due to the difficulty of local deployment, it is not easy to actually implement it.
[0003] Although cloud deployment is a shortcut, it can provide simple and convenient model deployment for small and medium-sized enterprises and users; however, there is a serious problem of privacy leakage, and a large number of enterprises are unwilling to share their data, which is difficult to implement

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
  • Deep learning model deployment method and device, and medium
  • Deep learning model deployment method and device, and medium
  • Deep learning model deployment method and device, and medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] Please refer to figure 1 and figure 2 , this embodiment mainly introduces a model training method on the Python side, by preprocessing the defect pictures obtained from the industrial camera and using the processed defect pictures to train the model;

[0056] Specifically, it includes the following steps:

[0057] S110. Defect image collection and defect labeling processing to generate a training data set.

[0058] S120. Expand the training data set; add defect types, and increase the training data set.

[0059] S130, dividing the training data set into training data and verification data;

[0060] Specifically, the training data set is randomly divided into training data and verification data in proportion;

[0061] Among them, the ratio of training data and validation data is 4:1.

[0062] S140. Select a model, train the model through the training data set, and adjust the training parameters according to the training effect;

[0063] Among them, the training pa...

Embodiment 2

[0072] Please refer to figure 1 and image 3 , this embodiment mainly introduces a Python-side model export method, including selecting a deployment framework type, exporting a model and a model description file; specifically, it includes the following steps:

[0073] S210. Select the deployment framework type;

[0074] S220, select a model format for the selected deployment framework type and then perform an export operation;

[0075] If you plan to deploy with the LibTorch framework, select PT for the model format in the export settings; if you plan to deploy with the TensorRT framework, select ONNX for the model format in the export settings;

[0076] S230, while exporting the model, generate a corresponding model description file;

[0077] The model description file contains: model task type (such as classification, positioning, segmentation or OCR), framework type (LibTorch, TensorRT), inference model type (accuracy priority, balance, speed priority), model loading ima...

Embodiment 3

[0080] Please refer to figure 1 , image 3 and Figure 4 , This embodiment mainly introduces a C++-side model part method, importing industrial pictures obtained from industrial cameras, automatically selecting the type of model deployment framework, and inferring and outputting detection results through model recognition: Specifically, it includes the following steps:

[0081] S310. Shooting through an industrial camera to obtain an industrial picture.

[0082] S320, load industrial pictures and model description files.

[0083] S330. Read the model description file, and automatically select corresponding inference parameters according to the model description file;

[0084] Specifically, read the model-related information (model task type, framework type, inference model type, model loading image length and width, and loading data arrangement format) in the model description file, and automatically set the corresponding inference parameters according to the model descript...

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 deep learning model deployment method and device, and a medium, and the method comprises the following steps: S100, model training: carrying out the preprocessing of a defect picture obtained from an industrial camera, and training a model through employing the processed defect picture; s200, model exporting: selecting a deployment framework type, and exporting a model and a model description file; s300, model deployment: importing an industrial picture obtained from an industrial camera, loading and reading a model description file, automatically selecting a corresponding inference parameter according to the model description file, and inferring and outputting a detection result through model identification; according to the method, the model description file is generated while the model is exported, the model description file is directly loaded at the model deployment end, the model inference frame type is automatically matched and selected, manual selection is not needed, and the operation is simple.

Description

technical field [0001] The invention relates to the technical field of industrial visual detection, and in particular to a deep learning model deployment method, equipment and medium. Background technique [0002] In recent years, the advantages of deep learning in computer vision have become more and more obvious, and industrial visual detection based on deep learning has also received more and more attention. However, due to the difficulty of local deployment, it is not easy to actually implement. [0003] Although cloud deployment is a shortcut, it can provide simple and convenient model deployment for small and medium-sized enterprises and users; however, there is a serious privacy leakage problem. [0004] For example, the published patent CN111488197A mainly provides simple and convenient model deployment for small and medium-sized enterprises and users, but relies too much on the cloud; the present invention proposes a deep learning model deployment method, which can ...

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): G06T7/00G06V10/96G06V10/94G06V10/774G06V10/82G06N3/08G06N3/10G06F8/60G06K9/62
CPCG06T7/0004G06N3/08G06N3/10G06F8/60G06T2207/20081G06T2207/20084G06T2207/30108G06F18/214
Inventor 强伟余章卫
Owner 苏州中科行智智能科技有限公司
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