sun-umn/PyGRANSO

PyGRANSO: A PyTorch-enabled port of GRANSO with auto-differentiation

56
/ 100
Established

PyGRANSO helps machine learning practitioners solve complex optimization problems where the objective function or constraints are 'nonsmooth' and 'nonconvex' — common challenges in deep learning. It takes your PyTorch-defined model and optimization problem, automatically handles the necessary calculus, and provides an optimized model that respects those difficult constraints. This is for anyone building advanced machine learning models, especially those in deep learning, who need to ensure their models meet specific criteria or performance bounds.

Available on PyPI.

Use this if you are a machine learning engineer or researcher working with PyTorch and need to train models under complex, nonsmooth, and nonconvex constraints, especially in deep learning applications.

Not ideal if your optimization problems are simple, smooth, convex, or if you prefer to manually provide analytical gradients for your optimization solver.

deep-learning machine-learning-optimization constrained-optimization neural-network-training model-regularization
No Dependents
Maintenance 10 / 25
Adoption 8 / 25
Maturity 25 / 25
Community 13 / 25

How are scores calculated?

Stars

42

Forks

6

Language

Python

License

AGPL-3.0

Last pushed

Feb 27, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/sun-umn/PyGRANSO"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.