# SimonBlanke/Gradient-Free-Optimizers

Lightweight optimization with local, global, population-based and sequential techniques across mixed search spaces

Repository: https://github.com/SimonBlanke/Gradient-Free-Optimizers
Canonical: https://ross.abutalabs.com/products/gradient-free-optimizers
Homepage: https://gradient-free-optimizers.readthedocs.io/en/latest/
Language: Python
License: MIT
License Family: permissive
Topics: optimization, gradient-free-optimization, bayesian-optimization, particle-swarm-optimization, evolution-strategies, simulated-annealing, random-search, hill-climbing, meta-heuristic, machine-learning, tree-of-parzen-estimator, hyperparameter-optimization, blackbox-optimization, constrained-optimization, nelder-mead
Last push: 2026-08-21T12:14:34+00:00

## Health v2 (maintenance only)
Score: 94/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 84, longevity 100
- inputs: {"age_days": 2342, "days_push": 12, "days_rel": 110, "gap_med": 24.5, "n_releases_24m": 11}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1273, forks 96 (observed 2026-08-28T04:04:12.296739+00:00)

## What it is
A Python library for gradient-free optimization of black-box functions, offering 23 algorithms (hill climbing, Bayesian optimization, particle swarm, simulated annealing, etc.) behind one unified API. It supports mixed search spaces combining continuous, discrete, and categorical dimensions, making it well suited to hyperparameter tuning and simulation optimization.

## Use cases
- tune machine learning model hyperparameters without gradients
- optimize a black-box objective function
- run bayesian optimization on a mixed search space
- compare metaheuristic optimizers like particle swarm and simulated annealing
- optimize simulation parameters where gradients are unavailable
- perform feature selection via search algorithms

## When to choose
- your objective function is non-differentiable, noisy, or expensive to evaluate
- you want to swap optimization algorithms with a one-line change
- your search space mixes continuous ranges, grids, and categorical choices
- you want a lightweight pure-Python optimizer with optional SciPy dependency

## When to avoid
- your problem has cheap, reliable gradients - use gradient-based optimizers like SciPy or PyTorch instead
- you need massively parallel or distributed optimization at scale
- you need constrained optimization with complex constraint solvers

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, search-engine, data-science
- domain: machine-learning, data-science, performance
- platform: python, cross-platform
- tags: optimization, hyperparameter-tuning, bayesian-optimization, metaheuristics, black-box-optimization, gradient-free, simulated-annealing, particle-swarm, evolution-strategies, hill-climbing, nelder-mead, search-space, algorithms

## Member repositories
- SimonBlanke/Gradient-Free-Optimizers (main) score 94

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:12.296739+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:03:11.289445+00:00, confidence not recorded.
  - readme: https://github.com/SimonBlanke/Gradient-Free-Optimizers (fetched 2026-08-28T04:04:12.296739+00:00, sha f2fee6fcec43)
  - registry_pypi: https://pypi.org/pypi/gradient-free-optimizers/json (fetched 2026-08-29T12:15:03.003796+00:00, sha 9ade8b5e0dd6)
- Data as of 2026-08-30T08:39:29.467469+00:00.