ray.tune.Tuner — Ray 2.45.0 (original) (raw)

class ray.tune.Tuner(trainable: str | Callable | Type[Trainable] | BaseTrainer | None = None, *, param_space: Dict[str, Any] | None = None, tune_config: TuneConfig | None = None, run_config: RunConfig | None = None, _tuner_kwargs: Dict | None = None, _tuner_internal: TunerInternal | None = None, _entrypoint: AirEntrypoint = AirEntrypoint.TUNER)[source]#

Tuner is the recommended way of launching hyperparameter tuning jobs with Ray Tune.

Parameters:

• trainable – The trainable to be tuned.
• param_space – Search space of the tuning job. See Working with Tune Search Spaces.
• tune_config – Tuning specific configs, such as setting customsearch algorithms andtrial scheduling algorithms.
• run_config – Job-level run configuration, which includes configs for persistent storage, checkpointing, fault tolerance, etc.

Usage pattern:

import ray.tune

def trainable(config): # Your training logic here ray.tune.report({"accuracy": 0.8})

tuner = Tuner( trainable=trainable, param_space={"lr": ray.tune.grid_search([0.001, 0.01])}, run_config=ray.tune.RunConfig(name="my_tune_run"), ) results = tuner.fit()

To retry a failed Tune run, you can then do

tuner = Tuner.restore(results.experiment_path, trainable=trainable) tuner.fit()

results.experiment_path can be retrieved from theResultGrid object. It can also be easily seen in the log output from your first run.

PublicAPI (beta): This API is in beta and may change before becoming stable.

Methods