optuna.integration.TensorFlowPruningHook¶
- class optuna.integration.TensorFlowPruningHook(trial: optuna.trial.Trial, estimator: tf.estimator.Estimator, metric: str, run_every_steps: int)[source]¶
TensorFlow SessionRunHook to prune unpromising trials.
See the example if you want to add a pruning hook to TensorFlow’s estimator.
- Parameters
trial – A
Trial
corresponding to the current evaluation of the objective function.estimator – An estimator which you will use.
metric – An evaluation metric for pruning, e.g.,
accuracy
andloss
.run_every_steps – An interval to watch the summary file.
- __init__(trial: optuna.trial.Trial, estimator: tf.estimator.Estimator, metric: str, run_every_steps: int) None [source]¶
Methods
__init__
(trial, estimator, metric, ...)after_run
(run_context, run_values)before_run
(run_context)begin
()