The project aims to create an advanced system supporting data annotation in the field of CV. The main functionalities include: creating new projects or continuing work on existing ones, defining data categories, adding photos, annotating visual data, managing annotations, selecting model architecture, and automatic training after manually tagging photos. Additionally, the system enables applying predictions to unlabeled data, correcting model results, analyzing data and versioning models.
- Label data manually (Label Studio)
- Train ML model for auto-labeling
- Check the results your model achieves
- Predict annotations with object detector
- Correct the predicted annotations manually (Label Studio)
- Download annotations in JSON file
- Save/download trained ML model
- Python
- Django
- Django Rest Framework
- React.js
- PyTorch
See the documentation for more detailed information: https://e-motion.readthedocs.io/en/latest/
To automatically install all app dependencies:
./install.shTo run (installed) application:
./run.shIt is recommended to use a virtual environment. If you selected the localization and the environment:
pip install -r requirements.txt
Download the image:
docker pull heartexlabs/label-studio:latest
On first run (create a container):
docker run --name label_studio --hostname=6c1add37c0c9 --user=1001 --env=PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin --env=DEBIAN_FRONTEND=noninteractive --env=LS_DIR=/label-studio --env=PIP_CACHE_DIR=/.cache --env=POETRY_CACHE_DIR=/.poetry-cache --env=DJANGO_SETTINGS_MODULE=core.settings.label_studio --env=LABEL_STUDIO_BASE_DATA_DIR=/label-studio/data --env=OPT_DIR=/opt/heartex/instance-data/etc --env=SETUPTOOLS_USE_DISTUTILS=stdlib --env=HOME=/label-studio --workdir=/label-studio -p 8089:8080 --label='org.opencontainers.image.ref.name=ubuntu' --label='org.opencontainers.image.version=22.04' --runtime=runc -d heartexlabs/label-studio:latest
To stop the container:
docker stop label_studio
On next runs (start existing container):
docker start label_studio
Please check the MIT license that is listed in this repository.