🚀 v1.0.0 - Release of Hungarian Network

Release v1.0.0

We are excited to announce the first official release of the Hungarian Network (Hnet) repository! This initial version lays the foundation for multi-source localization (DoA estimation) using the deep learning-based Hungarian algorithm implementation. Below are the key features, functionalities, and components included in this release.

🔑 Key Features

1. Synthetic Data Generation

• generate_hnet_training_data.py
  • Generates synthetic distance matrices and association matrices essential for training the Hnet model.
  • Supports various configurations for angular resolutions and DoA combinations.

2. Model Training

• run.py
  • Main script to train the Hnet model using the generated datasets.
  • Utilizes PyTorch Lightning for streamlined training loops and Hydra for configuration management.

3. Configuration Management

• configs/ Directory
  • Contains YAML configuration files.
  • Facilitates easy customization of training parameters, logging settings, and more.

4. Testing and Coverage

• test.py
  • Comprehensive test suite to validate model functionality and performance.
• pytest.ini
  • Configuration file for Pytest to manage test settings.
  • tests/ Directory : consistency and scenario tests are located in tests/ directory.
• htmlcov/ Directory
  • Contains HTML coverage reports to monitor test coverage and identify untested code segments.

6. Continuous Integration (CI)

• /.github/workflows/ci-cd.yml
  • GitHub Actions workflow file to automate testing, coverage checks, and Docker image builds upon commits and pull requests.