FREEDOM251

Freedom251 › ML Experiment Tracking Software

Oríon

Paid APILinuxMacSelf-hostedWindows

Visit Oríon Alternatives

Oríon's own site

What Oríon is

Oríon is a free, open-source asynchronous framework for black-box optimization of machine-learning models and training. It can tune command-line scripts with limited workflow changes: define search spaces, configure runs through arguments or files, and report objectives through its Python client API. Included algorithms are Random Search, Hyperband, TPE, and ASHA, with more available through plugins. Asynchronous parallel optimization supports Joblib and Dask workers. Trials and parameters can be versioned for reproducibility, with storage options including in-memory EphemeralDB, file-based PickledDB, and MongoDB. PickledDB is generally not suited to parallel optimization above 50 workers; MongoDB is recommended beyond that scale. Experiment views include regret curves, parallel coordinates, local parameter importance, and partial dependencies. Oríon runs on Linux, macOS, and Windows, installs from PyPI, and offers API and SDK access. It is self-hosted and licensed under BSD.

Who it suits

It suits machine-learning researchers and engineers tuning models or training through existing scripts. It is not itself a machine-learning framework, pipeline, or automated machine-learning product.

What it does well

  • Supports asynchronous parallel optimization
  • Includes four optimization algorithms
  • Joblib and Dask worker backends
  • Experiment versioning for reproducibility
  • Multiple database backends

What to know first

  • PickledDB is unsuitable above 50 workers
  • Additional algorithms require plugins
  • Self-hosted deployment required

Limits

PickledDB is generally not suited to parallel optimization with more than 50 workers; MongoDB is recommended above that scale. Oríon itself does not provide a machine learning framework, pipeline, or automated machine learning product.

Getting started

  1. Install Oríon from PyPI with pip install orion
  2. Define the parameter search space
  3. Configure the experiment with command-line arguments or a configuration file
  4. Add report_objective() or use the Python client API
  5. Choose EphemeralDB, PickledDB, or MongoDB storage
  6. Run workers with the documented Joblib or Dask backends

Verdict

Oríon adds configurable black-box optimization and experiment tracking to existing research workflows. Its algorithms, APIs, visualizations, and storage choices are useful, but larger parallel runs favor MongoDB over PickledDB.

Questions

Is Oríon free and open source?

Yes. Oríon is free, open source, and documented under the BSD license.

Which operating systems does it support?

It runs on Linux, macOS, and Windows, with API and SDK access.

How is Oríon installed?

Install it from PyPI using pip install orion.

Which algorithms are included?

The included algorithms are Random Search, Hyperband, TPE, and ASHA. Plugins can add more algorithms.

Which databases can store experiments?

The documented backends are in-memory EphemeralDB, file-based PickledDB, and MongoDB.

Can it run optimization in parallel?

Yes. Oríon supports asynchronous parallel optimization and documents Joblib and Dask worker backends.

Facts from Oríon's own pages

Algorithms
The installation guide lists Random Search, Hyperband, TPE, and ASHA as included algorithms, with more available through plugins. · orion.readthedocs.io, 4 Oct 2026
Database backends
The documented built-in backends are EphemeralDB, file-based PickledDB, and MongoDB; MongoDB is recommended for large-scale parallel optimization above 50 workers. · orion.readthedocs.io, 11 Oct 2026
Experiment tracking
Oríon stores trials in a configured database and offers experiment versioning to track trials and their parameters for reproducibility. · orion.readthedocs.io, 11 Oct 2026
Installation
The installation guide directs users to install Oríon from PyPI with pip install orion. · orion.readthedocs.io, 11 Oct 2026
Integration
A script can report its optimization objective through Oríon’s Python client API. · orion.readthedocs.io, 4 Oct 2026
Intended users
The documentation describes Oríon as useful for machine learning researchers and engineers, while saying it is not itself a machine learning framework, pipeline, or automated machine learning product. · orion.readthedocs.io, 11 Oct 2026
License
The project documentation says Oríon is licensed under the BSD license. · orion.readthedocs.io, 4 Oct 2026
Optimization algorithms
The documentation lists Random Search, Hyperband, TPE, and ASHA as included algorithms, and supports adding algorithms through plugins. · orion.readthedocs.io, 11 Oct 2026
Parallel execution
Oríon is natively asynchronous and the documentation says it can run in parallel from a laptop to a computing farm with thousands of processors. · orion.readthedocs.io, 11 Oct 2026
Platforms
The installation guide says Oríon supports most Linux distributions and Mac OS X, and Windows 10 through Windows Subsystem for Linux. · orion.readthedocs.io, 11 Oct 2026
Purpose
Oríon is an asynchronous framework for black-box function optimization, including hyperparameter optimization for machine learning. · orion.readthedocs.io, 11 Oct 2026
Scaling limit
The documentation says PickledDB is generally not suited to parallel optimization with more than 50 workers, though it describes this as a rule of thumb. · orion.readthedocs.io, 11 Oct 2026
Search spaces
Oríon supports search spaces defined by priors, including almost all SciPy distributions, configured through the command line or a configuration file. · orion.readthedocs.io, 11 Oct 2026
Storage
The documented database backends are in-memory EphemeralDB, file-based PickledDB, and MongoDB. · orion.readthedocs.io, 4 Oct 2026
Storage limit
The documentation says PickledDB is generally not suited to parallel optimization with more than 50 workers; MongoDB is recommended above that scale. · orion.readthedocs.io, 4 Oct 2026
Support
The project directs users to report questions, issues, bugs, or feature suggestions through its GitHub issue tracker. · orion.readthedocs.io, 11 Oct 2026
Visualization
The documentation includes experiment visualizations such as regret curves, parallel coordinates, local parameter importance, and partial dependencies. · orion.readthedocs.io, 4 Oct 2026
Workflow
It can tune command-line scripts with minimal changes to a researcher’s existing workflow. · orion.readthedocs.io, 4 Oct 2026
Workflow integration
The documentation says users can integrate Oríon into an existing workflow by calling report_objective() at the end of their script, or use its Python API. · orion.readthedocs.io, 11 Oct 2026

Where Oríon stands on Freedom251

Look at these too