AI Alliance – Efficient Labelling Service

This repository contains the Efficient Labelling Service developed in the AI Alliance project.
The framework provides:

• A Graphical User Interface (GUI) (built with Streamlit) for demonstration and exploration.
• A Service module containing the Similarity Pipeline and Labelling Pipeline, which can be used independently or in combination.
• A flexible architecture allowing you to use the full demonstrator, or integrate individual pipelines into your own workflow.

---

Repository Structure

.
├── datasets                  # Not contained in repository!!! Please download and put it here!
├── model_weights             # Not contained in repository!!! Please download and put it here!
├── gui/                      # Streamlit GUI demonstrator
├── service/                  # Core pipelines and service logic
│   ├── similarity_component/ # Similarity pipeline
│   ├── labelling_component/  # Labelling pipeline
│   └── main.py               # Entry point for combined service
├── assets/                   # Showcase datasets and images
└── README.md                 # (this file)

---

Quickstart

1. Clone the repository

git clone https://github.com/your-org/efficient-labelling-service.git
cd efficient-labelling-service

2. Install dependencies

pip install -r requirements.txt

3. Download the necessary folders

These folders need to be downloaded and put inside the top-level structure (see above).

.
├── datasets                  # Not contained in repository!!! Please download and put it here!
├── model_weights             # Not contained in repository!!! Please download and put it here!
...

You can download them here: Google Drive Link

4. Run the GUI

cd gui
streamlit run app.py

5. Use the pipelines directly

• Similarity Pipeline
• Labelling Pipeline

---

Docker Setup

Build and start the docker

docker compose up --build
docker compose up -d 

Terminate the docker

docker compose down 

---

Notes

• The datasets and model_weights folders are excluded from this repository. Please request access from the AI Alliance team.
• The GUI comes with showcase datasets for demonstration (/gui/assets/hannover_messe_data).

---

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

MIT License