optuna and orion

Optuna is a mature, general-purpose hyperparameter optimization framework that serves as a direct competitor to Orion, though Orion specializes in asynchronous distributed optimization with a focus on MongoDB-backed trial storage and experiment tracking, making it a niche alternative rather than a complement.

optuna
85
Verified
orion
68
Established
Maintenance 25/25
Adoption 15/25
Maturity 25/25
Community 20/25
Maintenance 6/25
Adoption 17/25
Maturity 25/25
Community 20/25
Stars: 13,673
Forks: 1,274
Downloads:
Commits (30d): 193
Language: Python
License: MIT
Stars: 301
Forks: 51
Downloads: 683
Commits (30d): 0
Language: Python
License:
No risk flags
No Dependents

About optuna

optuna/optuna

A hyperparameter optimization framework

Supports dynamic, conditionally-constructed search spaces through a define-by-run imperative API that allows hyperparameter dependencies. Leverages state-of-the-art sampling algorithms (TPE, GP-based, multi-objective optimization) with built-in early stopping via trial pruning. Integrates with PyTorch, TensorFlow/Keras, LightGBM, XGBoost, and distributed frameworks like Dask for scaling across multiple workers.

About orion

Epistimio/orion

Asynchronous Distributed Hyperparameter Optimization.

Scores updated daily from GitHub, PyPI, and npm data. How scores work