GENO is a solver for non-linear optimization problems. It can solve constrained and unconstrained problems. It is fully written in Python with no dependencies and it can run on the CPU and on the GPU.
pip install genosolver
See geno-project.org for more details, examples, and for an easy-to-read modeling language interface.
- Sören Laue - most of the solver -
- Mark Blacher - GPU support
- Matthias Mitterreiter
This project is licensed under the GNU Affero General Public License v3.
For more information, see the papers
- Optimization for Classical Machine Learning Problems on the GPU. Sören Laue, Mark Blacher, and Joachim Giesen. In AAAI Conference on Artificial Intelligence (AAAI), 2022.
- GENO -- GENeric Optimization for Classical Machine Learning. Sören Laue, Matthias Mitterreiter, and Joachim Giesen. In Advances in Neural Information Processing Systems (NeurIPS), 2019.