Ray Tune

E438346

Ray Tune is a scalable hyperparameter tuning and experiment management library for machine learning, built on the Ray distributed computing framework.

All labels observed (1)

Label Occurrences
Ray Tune canonical 1

How this entity was disambiguated

Statements (48)

Predicate Object
instanceOf Python library
experiment management library
hyperparameter optimization library
open-source software
basedOn Ray distributed computing framework NERFINISHED
developer Anyscale NERFINISHED
Ray open source community
documentation https://docs.ray.io/en/latest/tune/index.html
feature callbacks system
integration with Ray AIR
integration with Ray Train
scheduler abstraction
search algorithm abstraction
search space definition API
license Apache License 2.0
partOf Ray NERFINISHED
programmingLanguage Python
repository https://github.com/ray-project/ray
supports ASHA NERFINISHED
Bayesian optimization
HyperBand NERFINISHED
asynchronous hyperparameter optimization
cluster execution
distributed hyperparameter search
early stopping
experiment tracking
grid search
integration with Keras
integration with LightGBM
integration with PyTorch
integration with TensorFlow
integration with XGBoost
multi-GPU training
multi-node execution
parallel hyperparameter tuning
population-based training
random search
result logging
search algorithms from HyperOpt NERFINISHED
search algorithms from Nevergrad
search algorithms from Optuna
search algorithms from Scikit-Optimize NERFINISHED
synchronous hyperparameter optimization
trial checkpointing
useCase automated model optimization
hyperparameter tuning at scale
large-scale experiment management
machine learning model selection

How these facts were elicited

Referenced by (1)

Full triples — surface form annotated when it differs from this entity's canonical label.

RLlib integratesWith Ray Tune