History#

This page documents the history and evolution of Hyperactive.

Project History#

Hyperactive was created in 2018 by Simon Blanke to address the need for a flexible, unified interface for hyperparameter optimization in machine learning workflows.

Timeline#

2018 - Project Creation

Hyperactive was first released as an open-source project, providing a collection of gradient-free optimization algorithms accessible through a simple Python API.

2019 - Growing Adoption

The project gained traction in the machine learning community, with users appreciating its straightforward interface and variety of optimization algorithms.

2020-2021 - Ecosystem Expansion

Related projects were developed to complement Hyperactive:

  • Gradient-Free-Optimizers: The optimization backend was extracted into its own package, allowing for more modular development.

  • Search-Data-Collector: Tools for saving optimization results.

  • Search-Data-Explorer: Visualization dashboard for exploring search data.

2022-2023 - Continued Development

Active maintenance continued with bug fixes, new algorithms, and improved documentation. The user base continued to grow.

2024 - Version 5.0 Redesign

Major architecture redesign introducing:

  • Experiment-based architecture: Clean separation between optimization problems (experiments) and optimization algorithms (optimizers).

  • Enhanced integrations: Improved support for scikit-learn, sktime, skpro, and PyTorch Lightning.

  • Optuna backend: Integration with Optuna’s optimization algorithms.

  • Modern Python support: Support for Python 3.10 through 3.14.

2024 - GC.OS Sponsorship

Hyperactive became a sponsored project of the German Center for Open Source AI (GC.OS), ensuring continued development and maintenance.

Version History#

Major Versions#

Version

Highlights

v5.0

Experiment-based architecture, Optuna integration, modern Python support

v4.x

Improved API stability, additional optimizers

v3.x

Search data collection features, expanded algorithm library

v2.x

Multi-processing support, warm starting

v1.x

Initial public release with core optimization algorithms

Breaking Changes#

Major version updates (e.g., v4 → v5) may include breaking API changes. If you’re upgrading from an older version:

  1. Check the GitHub releases for migration guides.

  2. Update your code to use the new API patterns.

  3. Alternatively, pin your version to continue using the old API.

# Upgrade to latest
pip install hyperactive --upgrade

# Or pin to specific version
pip install hyperactive==4.x.x

Legacy Documentation#

Documentation for Hyperactive v4 is still available at the legacy documentation site:

Legacy Documentation (v4)

This may be useful if you:

  • Are maintaining projects that use Hyperactive v4

  • Need to reference the previous API design

  • Want to compare the old and new approaches

Future Roadmap#

Hyperactive continues to evolve. Planned improvements include:

  • Additional optimization algorithms

  • Enhanced visualization tools

  • Improved distributed computing support

  • More framework integrations

  • Performance optimizations

For the latest roadmap, see the GitHub Issues and Discussions.