AutomatedML

AutomatedML is a modular framework for automated machine learning, supporting neural network architectures, model evaluation, optimization, and data handling.

If you use this work in a paper, please cite the following publication:

A General-Purpose Neural Architecture Search Algorithm for Building Deep Neural Networks https://doi.org/10.1007/978-3-031-62922-8_9

Related publication:

Metaheuristics in Automated Machine Learning: Strategies for Optimization https://doi.org/10.1016/j.iswa.2025.200532

Features

• Modular component factory for ML pipelines
• Data container utilities
• Model evaluation and optimization
• Support for custom layers and models
• Extensible engine and utility modules

Directory Structure

src/automatedML/      # Core library modules
    ann/              # Neural network components
    annmodels/        # Predefined models
    component/        # Pipeline components
    engine/           # Execution engine
    flatgenerator/    # Model generator
    models/           # Model definitions
    optimizator/      # Optimization algorithms
    utils/            # Utility functions

tests/                # Test scripts and datasets
anns/                 # Example neural network configs and images
datasets/             # Example datasets
plots/                # Generated plots and visualizations

License

See LICENSE for details.