YOLOv8 interactive ML-assisted labeling, facilitating faster annotation for image detection, instance image segmentation.
Tested against Label Studio 1.13.1.
-
Dockerfile: The Dockerfile for building the backend container.
-
docker-compose.yml: The docker-compose file for running the backend.
-
_wsgi.py: WSGI app initializer.
-
start.sh: bash script to start the whole process.
-
model.py: The Python code for the ML backend model.
-
requirements.txt: The list of Python dependencies for the backend.
Before you begin:
- Ensure git is installed
- Ensure Docker Compose is installed.
Launch Label Studio. You can follow the guide from the official documentation or use the following commands:
If you're using local file serving, be sure to get a copy of the API token from Label Studio to connect the model.
Create a new project.
In the project Settings set up the Labeling Interface for image detection (RectangleLabels) or image segmentation (PolygonLabels).
Download the Label Studio YOLOv8 backend repository.
git clone https://github.com/seblful/label-studio-yolov8-backend.git
cd label-studio-yolov8-backend
Configure parameters in .env
file:
LABEL_STUDIO_URL=<IPv4 Address> (check your ipconfig)
LABEL_STUDIO_API_KEY=<Label Studio API token>
TASK_TYPE=<segmentation> or <detection>
docker compose up
Upload images directly to Label Studio using the Label Studio interface.
From the project settings, select the Model page and click Connect Model.
Add the URL http://locallhost:9090
and save the model as an ML backend.
To use this functionality, activate Auto-Annotation for drawing boxes.
Model training is not included in this project. This will probably be added later.
Contributions to this project are welcome. To contribute, please submit an issue or pull request.