X-AnyLabeling Deployment

Quick Start

• add the raw data folder path in docker-compose.yml

- type: bind # Bind mount the host directory into the container
    source: /home/yichen/datasets
    target: /root/datasets

• run the following command to start the docker container

git clone --recursive https://github.com/Ea510chan/anylabeling-deploy.git
cd anylabeling-deploy/docker && docker-compose up

(Optional) Enable CUDA-based Inference

• Change to GPU: et the preferred_device field to GPU in the app_info.py configuration file.

(Optional) Export LabelMe Annotations to VOC Format

• Export LabelMe Annotations to VOC Format Use the following command to convert LabelMe annotations to VOC format:

  python3 anylabeling-deploy/labelme/examples/semantic_segmentation/labelme2voc.py path/to/labelme/data path/to/output/voc --labels path/to/labels.txt --noobject

Acknowledgments

This project builds upon the excellent work done by the X-Anylabeling and AnyLabeling team. I'm grateful for their efforts in creating a flexible and powerful labeling tool for the community.