Tune Stopping Mechanisms (tune.stopper)#

In addition to Trial Schedulers like ASHA, where a number of trials are stopped if they perform subpar, Ray Tune also supports custom stopping mechanisms to stop trials early. They can also stop the entire experiment after a condition is met. For instance, stopping mechanisms can specify to stop trials when they reached a plateau and the metric doesn’t change anymore.

Ray Tune comes with several stopping mechanisms out of the box. For custom stopping behavior, you can inherit from the Stopper class.

Other stopping behaviors are described in the user guide.

Stopper Interface (tune.Stopper)#


Base class for implementing a Tune experiment stopper.

Stopper.__call__(trial_id, result)

Returns true if the trial should be terminated given the result.


Returns true if the experiment should be terminated.

Tune Built-in Stoppers#


Stop trials after reaching a maximum number of iterations

ExperimentPlateauStopper(metric[, std, top, ...])

Early stop the experiment when a metric plateaued across trials.

TrialPlateauStopper(metric[, std, ...])

Early stop single trials when they reached a plateau.


Stops all trials after a certain timeout.


Combine several stoppers via 'OR'.