def optimize_by_continuous_asha(
objective: Any,
dataset: TuneDataset,
plan: List[Tuple[float, int]],
checkpoint_path: str = "",
always_checkpoint: bool = False,
study_early_stop: Optional[Callable[[List[Any], List[RungHeap]], bool]] = None,
trial_early_stop: Optional[
Callable[[TrialReport, List[TrialReport], List[RungHeap]], bool]
] = None,
monitor: Any = None,
) -> StudyResult:
_objective = TUNE_OBJECT_FACTORY.make_iterative_objective(objective)
_monitor = TUNE_OBJECT_FACTORY.make_monitor(monitor)
checkpoint_path = TUNE_OBJECT_FACTORY.get_path_or_temp(checkpoint_path)
judge = ASHAJudge(
schedule=plan,
always_checkpoint=always_checkpoint,
study_early_stop=study_early_stop,
trial_early_stop=trial_early_stop,
monitor=_monitor,
)
path = os.path.join(checkpoint_path, str(uuid4()))
FileSystem().makedirs(path, recreate=True)
study = IterativeStudy(_objective, checkpoint_path=path)
return study.optimize(dataset, judge=judge)
def optimize_noniterative(
objective: Any,
dataset: TuneDataset,
runner: Any = None,
distributed: Optional[bool] = None,
monitor: Any = None,
) -> StudyResult:
_objective = TUNE_OBJECT_FACTORY.make_noniterative_objective(objective)
_runner = TUNE_OBJECT_FACTORY.make_noniterative_objective_runner(runner)
_monitor = TUNE_OBJECT_FACTORY.make_monitor(monitor)
study = NonIterativeStudy(_objective, _runner)
return study.optimize(dataset, distributed=distributed, monitor=_monitor)
def optimize_by_hyperband(
objective: Any,
dataset: TuneDataset,
plans: List[List[Tuple[float, int]]],
checkpoint_path: str = "",
distributed: Optional[bool] = None,
monitor: Any = None,
) -> StudyResult:
_monitor = TUNE_OBJECT_FACTORY.make_monitor(monitor)
weights = [float(p[0][1]) for p in plans]
datasets = dataset.divide(weights, seed=0)
result: Any = None
for d, plan in zip(datasets, plans):
r = optimize_by_sha(
objective=objective,
dataset=d,
plan=plan,
checkpoint_path=checkpoint_path,
distributed=distributed,
monitor=_monitor,
)
if result is None:
result = r
else:
result.union_with(r)
return result
def optimize_by_sha(
objective: Any,
dataset: TuneDataset,
plan: List[Tuple[float, int]],
checkpoint_path: str = "",
distributed: Optional[bool] = None,
monitor: Any = None,
) -> StudyResult:
_objective = TUNE_OBJECT_FACTORY.make_iterative_objective(objective)
_monitor = TUNE_OBJECT_FACTORY.make_monitor(monitor)
checkpoint_path = TUNE_OBJECT_FACTORY.get_path_or_temp(checkpoint_path)
path = os.path.join(checkpoint_path, str(uuid4()))
for budget, keep in plan:
obj = _NonIterativeObjectiveWrapper(
_objective, checkpoint_path=path, budget=budget
)
result = optimize_noniterative(
obj, dataset, distributed=distributed, monitor=_monitor
)
dataset = result.next_tune_dataset(keep)
return result
def suggest_by_hyperband(
objective: Any,
space: Space,
plans: List[List[Tuple[float, int]]],
train_df: Any = None,
temp_path: str = "",
partition_keys: Optional[List[str]] = None,
top_n: int = 1,
monitor: Any = None,
distributed: Optional[bool] = None,
execution_engine: Any = None,
execution_engine_conf: Any = None,
) -> List[TrialReport]:
assert_or_throw(
not space.has_random_parameter,
TuneCompileError("space can't contain random parameters, "
"use sample method before calling this function"),
)
dag = FugueWorkflow()
dataset = TUNE_OBJECT_FACTORY.make_dataset(
dag,
space,
df=train_df,
partition_keys=partition_keys,
temp_path=temp_path,
)
study = optimize_by_hyperband(
objective=objective,
dataset=dataset,
plans=plans,
checkpoint_path=temp_path,
distributed=distributed,
monitor=monitor,
)
study.result(top_n).yield_dataframe_as("result")
rows = list(
dag.run(
execution_engine,
conf=execution_engine_conf,
)["result"].as_dict_iterable())
return [
TrialReport.from_jsondict(json.loads(r[TUNE_REPORT]))
for r in sorted(rows, key=lambda r: r[TUNE_REPORT_METRIC])
]
def __init__(
self,
scoring: Any,
feature_prefix: str = "",
label_col: str = "label",
checkpoint_path: Optional[str] = None,
) -> None:
super().__init__()
self._last_id = ""
self._model_type: Any = None
self._model_expr: str = ""
self._scoring = scoring
self._feature_prefix = feature_prefix
self._label_col = label_col
if checkpoint_path is None:
self._checkpoint_path = checkpoint_path
else:
self._checkpoint_path = TUNE_OBJECT_FACTORY.get_path_or_temp(
checkpoint_path
)