PaddleLabel aims to become an effective and flexible data annotation tool. There are three parts to this project. This repo contains backend implementation. PaddleLabel-Frontend contains the React/Antd frontend. PaddleLabel-ML contains the machine learning backend for automatic and interactive models.

Install

Installing in a new enviroment is not required but suggested.

conda create -n pplabel python=3.9
conda activate pplabel

pip

pip install paddlelabel
paddlelabel

paddlelabel is now running at http://127.0.0.1:17995

source

First clone this repo for backend code.

git clone https://github.com/PaddleCV-SIG/PaddleLabel

Then clone and build frontend

git clone https://github.com/PaddleCV-SIG/PaddleLabel-Frontend
cd PaddleLabel-Frontend
npm install -g yarn
yarn
npm run build
cd ..

The last step is to copy built frontend to

cd PaddleLabel
pip install -r requirements.txt
mkdir paddlelabel/static/
cp -r ../PaddleLabel-Frontend/dist/* paddlelabel/static/

python setup.py install

Run

After installation, run PaddleLabel from command line with

paddlelabel

PaddleLabel is now avaliable at http://127.0.0.1:17995

You can also choose to expose the service to lan. This way it’s possbile to run the service on a computer and annotate with a tablet.

paddlelabel --lan

Dataset Import/Export

PaddleLabel currently support image classification, object detection and image segmentation projects. Please refer to the Dataset File Structure Documentation for more details.

Release Notes

• 2022.5.31: v0.1.0 [1] Support image classification, detection and segmentations. [2] Interactive image segmentation with EISeg models

Contribute

Please refer to the Developers Guide for details on backend implementation.