Supercharge Your Innovation With Domain-Expert AI Agents!

Neural network interactive automatic training system and method

A neural network and interactive technology, applied in the field of neural network interactive automatic training system, can solve the problems of unable to restore breakpoint progress, difficult to intervene, no overall progress, etc.

Pending Publication Date: 2021-03-19
ORIENTAL MIND (WUHAN) COMPUTING TECH CO LTD
View PDF0 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

There is no overall progress in the training process of these methods, hyperparameter tuning and architecture search can go on indefinitely, and after the training is interrupted, the breakpoint progress is often not restored like a single model
This is too black box for production development to intervene

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
  • Neural network interactive automatic training system and method
  • Neural network interactive automatic training system and method
  • Neural network interactive automatic training system and method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0057] In this embodiment, a method flow for data labeling is provided, and the method flow can be executed by the data labeling component 200 of the neural network interactive automatic training system.

[0058] according to diagram 2-1 As shown, the method for data labeling in this embodiment may include the following steps S201-S203:

[0059] Step S201, the user stores the required training data set in the data storage component. The data storage component may be any file system, such as object storage. The data set storage needs to meet certain rules, for example, the data set used for a task is stored in a separate path and not mixed with other files.

[0060] Step S202, in response to the target task type selected by the user, determine the target data set directory corresponding to the target task type, provide a labeling service based on the interface corresponding to the task type for the target task type, and Datasets in the dataset directory for display;

[006...

Embodiment 2

[0066] In this embodiment, a method flow for automatic model training is provided, and the method flow can be executed by the model training component 300 of the neural network interactive automatic training system.

[0067] according to image 3 As shown, the method for automatic model training in this embodiment may include the following modules 310-360. The model automatic training method is used to automatically train a neural network model on a given task and data set (including annotation files). According to the method of the present invention, in addition to receiving the task type, data set directory, and model storage path, in order to achieve adaptive data update and strong feedback in the training process, the following information can also be received.

[0068] Preferably, in order to improve the controllability of the training process, the input of the automatic model training method may also include optional model evaluation indicators, upper training duration ...

Embodiment 3

[0123] In this embodiment, a method for deploying a neural network model is provided, which can be executed by the model deploying component 400 of the neural network interactive automatic training system.

[0124] according to Figure 4 As shown, the method for implementing neural network model deployment in this embodiment may include the following steps S401-S406:

[0125] Step S401, acquiring a neural network model (trained model) generated by automatic model training. On first deployment, the model file path is specified externally. For example, according to the user interaction component 500 of the neural network interactive automatic training system of the present invention, it is specified by the user. Specifically, the existence of the model may be a single file, or a series of files related to the reasoning of the model. The present invention does not limit the file format and quantity of models.

[0126] Step S402, judging whether it is the first time to acquire...

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 neural network interactive automatic training system and method. The system comprises a data storage assembly, a data annotation assembly, a model training assembly and a model deployment assembly. The data annotation assembly is used for annotating an original data set provided by a user to generate an annotation file of the original data set; the model training assemblyis used for automatically carrying out neural network training based on the annotation file and generating a trained model; the model deployment component is used for deploying the trained model as an online reasoning service. According to the invention, a user performs data annotation in the neural network interactive automatic training system, the annotation result can be used for model training, and the training result can be deployed as an online service by one key, so that a full-automatic process from original data to an inference service is realized.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, and in particular relates to a neural network interactive automatic training system and method. Background technique [0002] The existing neural network automatic learning methods are basically aimed at neural network training. They use methods such as hyperparameter tuning and neural network architecture search to train on common benchmark data sets, and report the evaluation results on these data sets. These methods are widely used in academic research, but have the following problems in practical production applications: [0003] The scope of automation is limited. Most of the existing neural network automatic learning is for training, and it can only be performed on several general benchmark data sets. This is because the data processing methods of different algorithms may not be consistent, and the labeling formats of different data sets are not the same. This makes it dif...

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): G06N3/04G06N3/08G06N5/04
CPCG06N3/04G06N3/08G06N5/04
Inventor 罗明宇徐驰林健
Owner ORIENTAL MIND (WUHAN) COMPUTING TECH 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