Installation#

Hyperactive can be installed via pip and supports Python 3.10 - 3.14.

Installing Hyperactive#

Basic Installation#

Install Hyperactive from PyPI using pip:

pip install hyperactive

This installs Hyperactive with its core dependencies, which is sufficient for most use cases including scikit-learn integration.

Installation with Extras#

For additional functionality, you can install optional extras:

# Full installation with all extras (Optuna, PyTorch Lightning, etc.)
pip install hyperactive[all_extras]

# Scikit-learn integration (included by default)
pip install hyperactive[sklearn-integration]

# Sktime/skpro integration for time series
pip install hyperactive[sktime-integration]

Development Installation#

To install Hyperactive for development (from source):

# Clone the repository
git clone https://github.com/SimonBlanke/Hyperactive.git
cd Hyperactive

# Install in development mode with test dependencies
pip install -e ".[test]"

# Or install with all development dependencies
pip install -e ".[test,docs]"

Dependencies#

Core Dependencies#

Hyperactive requires the following packages (automatically installed):

Package	Purpose
`numpy >= 1.18.1`	Numerical operations and array handling
`pandas < 3.0.0`	Data manipulation and results handling
`tqdm >= 4.48.0`	Progress bars during optimization
`gradient-free-optimizers >= 1.2.4`	Core optimization algorithms
`scikit-base < 1.0.0`	Base classes for sklearn-like interfaces
`scikit-learn < 1.8.0`	Machine learning integration

Optional Dependencies#

Depending on your use case, you may want to install additional packages:

Optuna Backend#

For Optuna-based optimizers (TPE, CMA-ES, NSGA-II, etc.):

pip install optuna

Or include it via the all_extras option:

pip install hyperactive[all_extras]

Time Series (sktime/skpro)#

For time series forecasting and probabilistic prediction:

pip install hyperactive[sktime-integration]

This installs sktime and skpro for time series optimization experiments.

Note

Sktime integration requires Python < 3.14 due to sktime’s current compatibility.

PyTorch Lightning#

For deep learning hyperparameter optimization:

pip install hyperactive[all_extras]
# or
pip install lightning

Note

PyTorch/Lightning requires Python < 3.14 for full compatibility.

Verifying Installation#

After installation, verify that Hyperactive is working correctly:

import hyperactive
print(f"Hyperactive version: {hyperactive.__version__}")

# Quick test
import numpy as np
from hyperactive.opt.gfo import HillClimbing


def objective(params):
    return -(params["x"] ** 2)


optimizer = HillClimbing(
    search_space={"x": np.arange(-5, 5, 0.1)},
    n_iter=10,
    experiment=objective,
)
best = optimizer.solve()
print(f"Test optimization successful: {best}")

Python Version Support#

Hyperactive officially supports Python 3.10, 3.11, 3.12, 3.13, 3.14.

Note

The supported Python versions are automatically extracted from the project’s pyproject.toml classifiers.

Some optional integrations (sktime, PyTorch) may have more restrictive Python version requirements. Check the specific package documentation for details.

Troubleshooting#

Common Installation Issues#

ImportError: No module named ‘gradient_free_optimizers’

This usually means the installation was incomplete. Try:

pip install --upgrade hyperactive

MemoryError during optimization

Sequential model-based optimizers (Bayesian, TPE) can use significant memory for large search spaces. Reduce your search space size or use a simpler optimizer like RandomSearch or HillClimbing.

Pickle errors with multiprocessing

Ensure all objects in your search space are serializable (no lambdas, closures, or bound methods). Use top-level functions and basic Python types.

For more help, see the GitHub Issues.

Using Older Versions#

If you need to use Hyperactive v4, you can install a specific version:

pip install hyperactive==4.8.0

Documentation for v4 is available at: Legacy Documentation (v4)