import logging
import os
from pathlib import Path
from typing import TYPE_CHECKING, Any, Callable, Dict, Optional, Type, Union
import pyarrow.fs
import ray
from ray.air._internal.usage import AirEntrypoint
from ray.air.config import RunConfig
from ray.air.util.node import _force_on_current_node
from ray.train._internal.storage import _exists_at_fs_path, get_fs_and_path
from ray.tune import ResumeConfig
from ray.tune.experimental.output import get_air_verbosity
from ray.tune.impl.tuner_internal import _TUNER_PKL, TunerInternal
from ray.tune.progress_reporter import (
_prepare_progress_reporter_for_ray_client,
_stream_client_output,
)
from ray.tune.result_grid import ResultGrid
from ray.tune.trainable import Trainable
from ray.tune.tune_config import TuneConfig
from ray.util import PublicAPI
logger = logging.getLogger(__name__)
if TYPE_CHECKING:
from ray.train.base_trainer import BaseTrainer
ClientActorHandle = Any
# try:
# # Breaks lint right now.
# from ray.util.client.common import ClientActorHandle
# except Exception:
# pass
# The magic key that is used when instantiating Tuner during resume.
_TUNER_INTERNAL = "_tuner_internal"
_SELF = "self"
[docs]
@PublicAPI(stability="beta")
class Tuner:
"""Tuner is the recommended way of launching hyperparameter tuning jobs with Ray Tune.
Args:
trainable: The trainable to be tuned.
param_space: Search space of the tuning job.
One thing to note is that both preprocessor and dataset can be tuned here.
tune_config: Tuning algorithm specific configs.
Refer to ray.tune.tune_config.TuneConfig for more info.
run_config: Runtime configuration that is specific to individual trials.
If passed, this will overwrite the run config passed to the Trainer,
if applicable. Refer to ray.train.RunConfig for more info.
Usage pattern:
.. code-block:: python
from sklearn.datasets import load_breast_cancer
from ray import tune
from ray.data import from_pandas
from ray.train import RunConfig, ScalingConfig
from ray.train.xgboost import XGBoostTrainer
from ray.tune.tuner import Tuner
def get_dataset():
data_raw = load_breast_cancer(as_frame=True)
dataset_df = data_raw["data"]
dataset_df["target"] = data_raw["target"]
dataset = from_pandas(dataset_df)
return dataset
trainer = XGBoostTrainer(
label_column="target",
params={},
datasets={"train": get_dataset()},
)
param_space = {
"scaling_config": ScalingConfig(
num_workers=tune.grid_search([2, 4]),
resources_per_worker={
"CPU": tune.grid_search([1, 2]),
},
),
# You can even grid search various datasets in Tune.
# "datasets": {
# "train": tune.grid_search(
# [ds1, ds2]
# ),
# },
"params": {
"objective": "binary:logistic",
"tree_method": "approx",
"eval_metric": ["logloss", "error"],
"eta": tune.loguniform(1e-4, 1e-1),
"subsample": tune.uniform(0.5, 1.0),
"max_depth": tune.randint(1, 9),
},
}
tuner = Tuner(trainable=trainer, param_space=param_space,
run_config=RunConfig(name="my_tune_run"))
results = tuner.fit()
To retry a failed tune run, you can then do
.. code-block:: python
tuner = Tuner.restore(results.experiment_path, trainable=trainer)
tuner.fit()
``results.experiment_path`` can be retrieved from the
:ref:`ResultGrid object <tune-analysis-docs>`. It can
also be easily seen in the log output from your first run.
"""
# One of the following is assigned.
_local_tuner: Optional[TunerInternal] # Only used in none ray client mode.
_remote_tuner: Optional[ClientActorHandle] # Only used in ray client mode.
[docs]
def __init__(
self,
trainable: Optional[
Union[str, Callable, Type[Trainable], "BaseTrainer"]
] = None,
*,
param_space: Optional[Dict[str, Any]] = None,
tune_config: Optional[TuneConfig] = None,
run_config: Optional[RunConfig] = None,
# This is internal only arg.
# Only for dogfooding purposes. We can slowly promote these args
# to RunConfig or TuneConfig as needed.
# TODO(xwjiang): Remove this later.
_tuner_kwargs: Optional[Dict] = None,
_tuner_internal: Optional[TunerInternal] = None,
_entrypoint: AirEntrypoint = AirEntrypoint.TUNER,
):
"""Configure and construct a tune run."""
kwargs = locals().copy()
self._is_ray_client = ray.util.client.ray.is_connected()
if self._is_ray_client:
_run_config = run_config or RunConfig()
if get_air_verbosity(_run_config.verbose) is not None:
logger.info(
"[output] This uses the legacy output and progress reporter, "
"as Ray client is not supported by the new engine. "
"For more information, see "
"https://github.com/ray-project/ray/issues/36949"
)
if _tuner_internal:
if not self._is_ray_client:
self._local_tuner = kwargs[_TUNER_INTERNAL]
else:
self._remote_tuner = kwargs[_TUNER_INTERNAL]
else:
kwargs.pop(_TUNER_INTERNAL, None)
kwargs.pop(_SELF, None)
if not self._is_ray_client:
self._local_tuner = TunerInternal(**kwargs)
else:
self._remote_tuner = _force_on_current_node(
ray.remote(num_cpus=0)(TunerInternal)
).remote(**kwargs)
[docs]
@classmethod
def restore(
cls,
path: str,
trainable: Union[str, Callable, Type[Trainable], "BaseTrainer"],
resume_unfinished: bool = True,
resume_errored: bool = False,
restart_errored: bool = False,
param_space: Optional[Dict[str, Any]] = None,
storage_filesystem: Optional[pyarrow.fs.FileSystem] = None,
_resume_config: Optional[ResumeConfig] = None,
) -> "Tuner":
"""Restores Tuner after a previously failed run.
All trials from the existing run will be added to the result table. The
argument flags control how existing but unfinished or errored trials are
resumed.
Finished trials are always added to the overview table. They will not be
resumed.
Unfinished trials can be controlled with the ``resume_unfinished`` flag.
If ``True`` (default), they will be continued. If ``False``, they will
be added as terminated trials (even if they were only created and never
trained).
Errored trials can be controlled with the ``resume_errored`` and
``restart_errored`` flags. The former will resume errored trials from
their latest checkpoints. The latter will restart errored trials from
scratch and prevent loading their last checkpoints.
.. note::
Restoring an experiment from a path that's pointing to a *different*
location than the original experiment path is supported.
However, Ray Tune assumes that the full experiment directory is available
(including checkpoints) so that it's possible to resume trials from their
latest state.
For example, if the original experiment path was run locally,
then the results are uploaded to cloud storage, Ray Tune expects the full
contents to be available in cloud storage if attempting to resume
via ``Tuner.restore("s3://...")``. The restored run will continue
writing results to the same cloud storage location.
Args:
path: The local or remote path of the experiment directory
for an interrupted or failed run.
Note that an experiment where all trials finished will not be resumed.
This information could be easily located near the end of the
console output of previous run.
trainable: The trainable to use upon resuming the experiment.
This should be the same trainable that was used to initialize
the original Tuner.
param_space: The same `param_space` that was passed to
the original Tuner. This can be optionally re-specified due
to the `param_space` potentially containing Ray object
references (tuning over Datasets or tuning over
several `ray.put` object references). **Tune expects the
`param_space` to be unmodified**, and the only part that
will be used during restore are the updated object references.
Changing the hyperparameter search space then resuming is NOT
supported by this API.
resume_unfinished: If True, will continue to run unfinished trials.
resume_errored: If True, will re-schedule errored trials and try to
restore from their latest checkpoints.
restart_errored: If True, will re-schedule errored trials but force
restarting them from scratch (no checkpoint will be loaded).
storage_filesystem: Custom ``pyarrow.fs.FileSystem``
corresponding to the ``path``. This may be necessary if the original
experiment passed in a custom filesystem.
_resume_config: [Experimental] Config object that controls how to resume
trials of different statuses. Can be used as a substitute to
`resume_*` and `restart_*` flags above.
"""
unfinished = (
ResumeConfig.ResumeType.RESUME
if resume_unfinished
else ResumeConfig.ResumeType.SKIP
)
errored = ResumeConfig.ResumeType.SKIP
if resume_errored:
errored = ResumeConfig.ResumeType.RESUME
elif restart_errored:
errored = ResumeConfig.ResumeType.RESTART
resume_config = _resume_config or ResumeConfig(
unfinished=unfinished, errored=errored
)
if not ray.util.client.ray.is_connected():
tuner_internal = TunerInternal(
restore_path=path,
resume_config=resume_config,
trainable=trainable,
param_space=param_space,
storage_filesystem=storage_filesystem,
)
return Tuner(_tuner_internal=tuner_internal)
else:
tuner_internal = _force_on_current_node(
ray.remote(num_cpus=0)(TunerInternal)
).remote(
restore_path=path,
resume_config=resume_config,
trainable=trainable,
param_space=param_space,
storage_filesystem=storage_filesystem,
)
return Tuner(_tuner_internal=tuner_internal)
[docs]
@classmethod
def can_restore(
cls,
path: Union[str, os.PathLike],
storage_filesystem: Optional[pyarrow.fs.FileSystem] = None,
) -> bool:
"""Checks whether a given directory contains a restorable Tune experiment.
Usage Pattern:
Use this utility to switch between starting a new Tune experiment
and restoring when possible. This is useful for experiment fault-tolerance
when re-running a failed tuning script.
.. code-block:: python
import os
from ray.tune import Tuner
from ray.train import RunConfig
def train_fn(config):
# Make sure to implement checkpointing so that progress gets
# saved on restore.
pass
name = "exp_name"
storage_path = os.path.expanduser("~/ray_results")
exp_dir = os.path.join(storage_path, name)
if Tuner.can_restore(exp_dir):
tuner = Tuner.restore(exp_dir, trainable=train_fn, resume_errored=True)
else:
tuner = Tuner(
train_fn,
run_config=RunConfig(name=name, storage_path=storage_path),
)
tuner.fit()
Args:
path: The path to the experiment directory of the Tune experiment.
This can be either a local directory or a remote URI
(e.g. s3://bucket/exp_name).
Returns:
bool: True if this path exists and contains the Tuner state to resume from
"""
fs, fs_path = get_fs_and_path(path, storage_filesystem)
return _exists_at_fs_path(fs, Path(fs_path, _TUNER_PKL).as_posix())
def _prepare_remote_tuner_for_jupyter_progress_reporting(self):
run_config: RunConfig = ray.get(self._remote_tuner.get_run_config.remote())
progress_reporter, string_queue = _prepare_progress_reporter_for_ray_client(
run_config.progress_reporter, run_config.verbose
)
run_config.progress_reporter = progress_reporter
ray.get(
self._remote_tuner.set_run_config_and_remote_string_queue.remote(
run_config, string_queue
)
)
return progress_reporter, string_queue
[docs]
def fit(self) -> ResultGrid:
"""Executes hyperparameter tuning job as configured and returns result.
Failure handling:
For the kind of exception that happens during the execution of a trial,
one may inspect it together with stacktrace through the returned result grid.
See ``ResultGrid`` for reference. Each trial may fail up to a certain number.
This is configured by ``RunConfig.FailureConfig.max_failures``.
Exception that happens beyond trials will be thrown by this method as well.
In such cases, there will be instruction like the following printed out
at the end of console output to inform users on how to resume.
Please use `Tuner.restore` to resume.
.. code-block:: python
import os
from ray.tune import Tuner
trainable = ...
tuner = Tuner.restore(
os.path.expanduser("~/ray_results/tuner_resume"),
trainable=trainable
)
tuner.fit()
Raises:
RayTaskError: If user-provided trainable raises an exception
"""
if not self._is_ray_client:
return self._local_tuner.fit()
else:
(
progress_reporter,
string_queue,
) = self._prepare_remote_tuner_for_jupyter_progress_reporting()
fit_future = self._remote_tuner.fit.remote()
_stream_client_output(
fit_future,
progress_reporter,
string_queue,
)
return ray.get(fit_future)
[docs]
def get_results(self) -> ResultGrid:
"""Get results of a hyperparameter tuning run.
This method returns the same results as :meth:`~ray.tune.Tuner.fit`
and can be used to retrieve the results after restoring a tuner without
calling ``fit()`` again.
If the tuner has not been fit before, an error will be raised.
.. code-block:: python
from ray.tune import Tuner
# `trainable` is what was passed in to the original `Tuner`
tuner = Tuner.restore("/path/to/experiment', trainable=trainable)
results = tuner.get_results()
Returns:
Result grid of a previously fitted tuning run.
"""
if not self._is_ray_client:
return self._local_tuner.get_results()
else:
(
progress_reporter,
string_queue,
) = self._prepare_remote_tuner_for_jupyter_progress_reporting()
get_results_future = self._remote_tuner.get_results.remote()
_stream_client_output(
get_results_future,
progress_reporter,
string_queue,
)
return ray.get(get_results_future)
def __getattribute__(self, item):
if item == "restore":
raise AttributeError(
"`Tuner.restore()` is a classmethod and cannot be called on an "
"instance. Use `tuner = Tuner.restore(...)` to instantiate the "
"Tuner instead."
)
return super().__getattribute__(item)