(File-based) Journal Storage

Optuna provides JournalStorage. With this feature, you can easily run a distributed optimization over network using NFS as the shared storage, without need for setting up RDB or Redis.

import logging
import sys

import optuna


# Add stream handler of stdout to show the messages
optuna.logging.get_logger("optuna").addHandler(logging.StreamHandler(sys.stdout))
study_name = "example-study"  # Unique identifier of the study.
file_path = "./optuna_journal_storage.log"
storage = optuna.storages.JournalStorage(
    optuna.storages.journal.JournalFileBackend(file_path),  # NFS path for distributed optimization
)

study = optuna.create_study(study_name=study_name, storage=storage)


def objective(trial):
    x = trial.suggest_float("x", -10, 10)
    return (x - 2) ** 2


study.optimize(objective, n_trials=3)

A new study created in Journal with name: example-study
Trial 0 finished with value: 4.493782967283395 and parameters: {'x': -0.1198544684207441}. Best is trial 0 with value: 4.493782967283395.
Trial 1 finished with value: 44.69634210286845 and parameters: {'x': 8.685532297646049}. Best is trial 0 with value: 4.493782967283395.
Trial 2 finished with value: 22.897244778447543 and parameters: {'x': -2.785106558734878}. Best is trial 0 with value: 4.493782967283395.

Although the optimization in this example is too short to run in parallel, you can extend this example to write a optimization script which can be run in parallel.

Note

In a Windows environment, an error message “A required privilege is not held by the client” may appear. In this case, you can solve the problem with creating storage by specifying JournalFileOpenLock. See the reference of JournalStorage for any details.