Model Clustering and Consolidation System for Machine Learning Models
ModelForge is an advanced system for clustering and consolidating machine learning models. It efficiently reduces a large collection of models into a smaller set of representative models while preserving performance characteristics.
- Model Clustering - Group similar models based on performance metrics and characteristics
- Model Consolidation - Generate a smaller, representative set of models from the original collection
- Performance Evaluation - Comprehensively evaluate consolidated models against the original set
- Visualization Tools - Analyze model similarities and differences through intuitive visualizations
- Python 3.11 or higher
- Poetry (dependency management)
Clone the repository and install dependencies:
git clone https://github.com/yourusername/modelforge.git
cd modelforge
poetry installSee our demo notebook for a minimal example of how to use ModelForge.
- Using Your Own Datasets - Learn how to create and format your own datasets
- Evaluation Datasets - Details on datasets used in ModelForge evaluation
- Reproducibility - Instructions to reproduce our research results
- Visualization & Plots - Detailed explanation of visualization options and additional results
Our research demonstrates significant efficiency gains when using ModelForge for model consolidation while maintaining performance thresholds. See our paper for complete details.
If you use ModelForge in your research, please cite our paper:
@inproceedings{Siepe2025,
author={Siepe, Florian and Prinz, Thomas Peter and Papenbrock, Thorsten},
booktitle={2025 IEEE 12th International Conference on Data Science and Advanced Analytics (DSAA)},
title={ModelForge: A Metric Approach to Machine Learning Model Consolidation},
year={2025},
volume={},
number={},
pages={1-12},
doi={10.1109/DSAA65442.2025.11247992}
}Contributions are welcome! Please feel free to submit a Pull Request.
This project is licensed under the MIT License - see the LICENSE file for details.