English | 简体中文

Introduction

TextileAD-UI is an intuitive tool for training and inference in anomaly detection, designed to help users train and evaluate models without the need for coding. It leverages the anomalib backend for anomaly detection processing and uses vue-element-admin for the frontend user interface.

Getting started

System requirements

Python==3.10

npm==6.14.10

vue==2.6.10

Browsers support

Modern browsers and Internet Explorer 10+.

IE / Edge	Firefox	Chrome	Safari
IE10, IE11, Edge	last 2 versions	last 2 versions	last 2 versions

Server

# install dependency
pip install -r requirements.txt

# enter the core directory
cd core

# run
python app.py

# develop
python app.py --debug

This will automatically run on http://localhost:5000

Frontend

# clone the project
git clone https://github.com/BlueberryOreo/TextileAD-UI.git

# enter the project directory
cd TextileAD-UI

# install dependency
npm install

# develop
npm run dev

This will automatically open http://localhost:9527

Build

# build for test environment
npm run build:stage

# build for production environment
npm run build:prod

Advanced

# preview the release environment effect
npm run preview

# preview the release environment effect + static resource analysis
npm run preview -- --report

# code format check
npm run lint

# code format check and auto fix
npm run lint -- --fix

Refer to Documentation for more information about vue-element-admin

Snapshots

Train settings

Test settings

Training progress

Results

After training

After testing

Acknowledgements

This project is built on anomalib and vue-element-admin. The authors thank everyone who makes their code available.