ma921/SOBER

Fast Bayesian optimization, quadrature, inference over arbitrary domain with GPU parallel acceleration

35
/ 100
Emerging

This project helps scientists, engineers, and researchers efficiently find optimal solutions for complex problems with many variables, like in drug discovery or material design. You provide initial observations of your experimental results, and it recommends the next set of experiments (batch queries) to run, leading you to the best possible outcome faster and more reliably. It's for anyone needing to optimize 'black box' functions with limited experimental trials across diverse data types.

Use this if you need to quickly and accurately find the best settings or parameters for a system or experiment where each test is expensive or time-consuming, and you're working with continuous, discrete, or mixed types of input data.

Not ideal if your optimization problem has a simple, known mathematical form that can be solved directly without needing to learn from observations.

drug-discovery material-science experimental-design hyperparameter-tuning simulation-optimization
No Package No Dependents
Maintenance 6 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 6 / 25

How are scores calculated?

Stars

33

Forks

2

Language

Jupyter Notebook

License

BSD-3-Clause

Last pushed

Dec 10, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/ma921/SOBER"

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