397b843890
### What changes were proposed in this pull request? Code in the PR generates random parameters for hyperparameter tuning. A discussion with Sean Owen can be found on the dev mailing list here: http://apache-spark-developers-list.1001551.n3.nabble.com/Hyperparameter-Optimization-via-Randomization-td30629.html All code is entirely my own work and I license the work to the project under the project’s open source license. ### Why are the changes needed? Randomization can be a more effective techinique than a grid search since min/max points can fall between the grid and never be found. Randomisation is not so restricted although the probability of finding minima/maxima is dependent on the number of attempts. Alice Zheng has an accessible description on how this technique works at https://www.oreilly.com/library/view/evaluating-machine-learning/9781492048756/ch04.html Although there are Python libraries with more sophisticated techniques, not every Spark developer is using Python. ### Does this PR introduce _any_ user-facing change? A new class (`ParamRandomBuilder.scala`) and its tests have been created but there is no change to existing code. This class offers an alternative to `ParamGridBuilder` and can be dropped into the code wherever `ParamGridBuilder` appears. Indeed, it extends `ParamGridBuilder` and is completely compatible with its interface. It merely adds one method that provides a range over which a hyperparameter will be randomly defined. ### How was this patch tested? Tests `ParamRandomBuilderSuite.scala` and `RandomRangesSuite.scala` were added. `ParamRandomBuilderSuite` is the analogue of the already existing `ParamGridBuilderSuite` which tests the user-facing interface. `RandomRangesSuite` uses ScalaCheck to test the random ranges over which hyperparameters are distributed. Closes #31535 from PhillHenry/ParamRandomBuilder. Authored-by: Phillip Henry <PhillHenry@gmail.com> Signed-off-by: Sean Owen <srowen@gmail.com>
231 lines
8.7 KiB
Python
231 lines
8.7 KiB
Python
#
|
|
# Licensed to the Apache Software Foundation (ASF) under one
|
|
# or more contributor license agreements. See the NOTICE file
|
|
# distributed with this work for additional information
|
|
# regarding copyright ownership. The ASF licenses this file
|
|
# to you under the Apache License, Version 2.0 (the
|
|
# "License"); you may not use this file except in compliance
|
|
# with the License. You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing,
|
|
# software distributed under the License is distributed on an
|
|
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
|
# KIND, either express or implied. See the License for the
|
|
# specific language governing permissions and limitations
|
|
# under the License.
|
|
|
|
from typing import overload
|
|
from typing import Any, List, Optional, Tuple, Type
|
|
from pyspark.ml._typing import ParamMap
|
|
|
|
from pyspark.ml import Estimator, Model
|
|
from pyspark.ml.evaluation import Evaluator
|
|
from pyspark.ml.param import Param
|
|
from pyspark.ml.param.shared import HasCollectSubModels, HasParallelism, HasSeed
|
|
from pyspark.ml.util import MLReader, MLReadable, MLWriter, MLWritable
|
|
|
|
class ParamGridBuilder:
|
|
def __init__(self) -> None: ...
|
|
def addGrid(self, param: Param, values: List[Any]) -> ParamGridBuilder: ...
|
|
@overload
|
|
def baseOn(self, __args: ParamMap) -> ParamGridBuilder: ...
|
|
@overload
|
|
def baseOn(self, *args: Tuple[Param, Any]) -> ParamGridBuilder: ...
|
|
def build(self) -> List[ParamMap]: ...
|
|
|
|
class ParamRandomBuilder(ParamGridBuilder):
|
|
def __init__(self) -> None: ...
|
|
def addRandom(self, param: Param, x: Any, y: Any, n: int) -> ParamRandomBuilder: ...
|
|
def addLog10Random(self, param: Param, x: Any, y: Any, n: int) -> ParamRandomBuilder: ...
|
|
|
|
class _ValidatorParams(HasSeed):
|
|
estimator: Param[Estimator]
|
|
estimatorParamMaps: Param[List[ParamMap]]
|
|
evaluator: Param[Evaluator]
|
|
def getEstimator(self) -> Estimator: ...
|
|
def getEstimatorParamMaps(self) -> List[ParamMap]: ...
|
|
def getEvaluator(self) -> Evaluator: ...
|
|
|
|
class _CrossValidatorParams(_ValidatorParams):
|
|
numFolds: Param[int]
|
|
foldCol: Param[str]
|
|
def __init__(self, *args: Any): ...
|
|
def getNumFolds(self) -> int: ...
|
|
def getFoldCol(self) -> str: ...
|
|
|
|
class CrossValidator(
|
|
Estimator[CrossValidatorModel],
|
|
_CrossValidatorParams,
|
|
HasParallelism,
|
|
HasCollectSubModels,
|
|
MLReadable[CrossValidator],
|
|
MLWritable,
|
|
):
|
|
def __init__(
|
|
self,
|
|
*,
|
|
estimator: Optional[Estimator] = ...,
|
|
estimatorParamMaps: Optional[List[ParamMap]] = ...,
|
|
evaluator: Optional[Evaluator] = ...,
|
|
numFolds: int = ...,
|
|
seed: Optional[int] = ...,
|
|
parallelism: int = ...,
|
|
collectSubModels: bool = ...,
|
|
foldCol: str = ...
|
|
) -> None: ...
|
|
def setParams(
|
|
self,
|
|
*,
|
|
estimator: Optional[Estimator] = ...,
|
|
estimatorParamMaps: Optional[List[ParamMap]] = ...,
|
|
evaluator: Optional[Evaluator] = ...,
|
|
numFolds: int = ...,
|
|
seed: Optional[int] = ...,
|
|
parallelism: int = ...,
|
|
collectSubModels: bool = ...,
|
|
foldCol: str = ...
|
|
) -> CrossValidator: ...
|
|
def setEstimator(self, value: Estimator) -> CrossValidator: ...
|
|
def setEstimatorParamMaps(self, value: List[ParamMap]) -> CrossValidator: ...
|
|
def setEvaluator(self, value: Evaluator) -> CrossValidator: ...
|
|
def setNumFolds(self, value: int) -> CrossValidator: ...
|
|
def setFoldCol(self, value: str) -> CrossValidator: ...
|
|
def setSeed(self, value: int) -> CrossValidator: ...
|
|
def setParallelism(self, value: int) -> CrossValidator: ...
|
|
def setCollectSubModels(self, value: bool) -> CrossValidator: ...
|
|
def copy(self, extra: Optional[ParamMap] = ...) -> CrossValidator: ...
|
|
def write(self) -> MLWriter: ...
|
|
@classmethod
|
|
def read(cls: Type[CrossValidator]) -> MLReader: ...
|
|
|
|
class CrossValidatorModel(
|
|
Model, _CrossValidatorParams, MLReadable[CrossValidatorModel], MLWritable
|
|
):
|
|
bestModel: Model
|
|
avgMetrics: List[float]
|
|
subModels: List[List[Model]]
|
|
def __init__(
|
|
self,
|
|
bestModel: Model,
|
|
avgMetrics: Optional[List[float]] = ...,
|
|
subModels: Optional[List[List[Model]]] = ...,
|
|
) -> None: ...
|
|
def copy(self, extra: Optional[ParamMap] = ...) -> CrossValidatorModel: ...
|
|
def write(self) -> MLWriter: ...
|
|
@classmethod
|
|
def read(cls: Type[CrossValidatorModel]) -> MLReader: ...
|
|
|
|
class _TrainValidationSplitParams(_ValidatorParams):
|
|
trainRatio: Param[float]
|
|
def __init__(self, *args: Any): ...
|
|
def getTrainRatio(self) -> float: ...
|
|
|
|
class TrainValidationSplit(
|
|
Estimator[TrainValidationSplitModel],
|
|
_TrainValidationSplitParams,
|
|
HasParallelism,
|
|
HasCollectSubModels,
|
|
MLReadable[TrainValidationSplit],
|
|
MLWritable,
|
|
):
|
|
def __init__(
|
|
self,
|
|
*,
|
|
estimator: Optional[Estimator] = ...,
|
|
estimatorParamMaps: Optional[List[ParamMap]] = ...,
|
|
evaluator: Optional[Evaluator] = ...,
|
|
trainRatio: float = ...,
|
|
parallelism: int = ...,
|
|
collectSubModels: bool = ...,
|
|
seed: Optional[int] = ...
|
|
) -> None: ...
|
|
def setParams(
|
|
self,
|
|
*,
|
|
estimator: Optional[Estimator] = ...,
|
|
estimatorParamMaps: Optional[List[ParamMap]] = ...,
|
|
evaluator: Optional[Evaluator] = ...,
|
|
trainRatio: float = ...,
|
|
parallelism: int = ...,
|
|
collectSubModels: bool = ...,
|
|
seed: Optional[int] = ...
|
|
) -> TrainValidationSplit: ...
|
|
def setEstimator(self, value: Estimator) -> TrainValidationSplit: ...
|
|
def setEstimatorParamMaps(self, value: List[ParamMap]) -> TrainValidationSplit: ...
|
|
def setEvaluator(self, value: Evaluator) -> TrainValidationSplit: ...
|
|
def setTrainRatio(self, value: float) -> TrainValidationSplit: ...
|
|
def setSeed(self, value: int) -> TrainValidationSplit: ...
|
|
def setParallelism(self, value: int) -> TrainValidationSplit: ...
|
|
def setCollectSubModels(self, value: bool) -> TrainValidationSplit: ...
|
|
def copy(self, extra: Optional[ParamMap] = ...) -> TrainValidationSplit: ...
|
|
def write(self) -> MLWriter: ...
|
|
@classmethod
|
|
def read(cls: Type[TrainValidationSplit]) -> MLReader: ...
|
|
|
|
class TrainValidationSplitModel(
|
|
Model,
|
|
_TrainValidationSplitParams,
|
|
MLReadable[TrainValidationSplitModel],
|
|
MLWritable,
|
|
):
|
|
bestModel: Model
|
|
validationMetrics: List[float]
|
|
subModels: List[Model]
|
|
def __init__(
|
|
self,
|
|
bestModel: Model,
|
|
validationMetrics: Optional[List[float]] = ...,
|
|
subModels: Optional[List[Model]] = ...,
|
|
) -> None: ...
|
|
def setEstimator(self, value: Estimator) -> TrainValidationSplitModel: ...
|
|
def setEstimatorParamMaps(
|
|
self, value: List[ParamMap]
|
|
) -> TrainValidationSplitModel: ...
|
|
def setEvaluator(self, value: Evaluator) -> TrainValidationSplitModel: ...
|
|
def copy(self, extra: Optional[ParamMap] = ...) -> TrainValidationSplitModel: ...
|
|
def write(self) -> MLWriter: ...
|
|
@classmethod
|
|
def read(cls: Type[TrainValidationSplitModel]) -> MLReader: ...
|
|
|
|
class CrossValidatorWriter(MLWriter):
|
|
instance: CrossValidator
|
|
def __init__(self, instance: CrossValidator) -> None: ...
|
|
def saveImpl(self, path: str) -> None: ...
|
|
|
|
class CrossValidatorReader(MLReader[CrossValidator]):
|
|
cls: Type[CrossValidator]
|
|
def __init__(self, cls: Type[CrossValidator]) -> None: ...
|
|
def load(self, path: str) -> CrossValidator: ...
|
|
|
|
class CrossValidatorModelWriter(MLWriter):
|
|
instance: CrossValidatorModel
|
|
def __init__(self, instance: CrossValidatorModel) -> None: ...
|
|
def saveImpl(self, path: str) -> None: ...
|
|
|
|
class CrossValidatorModelReader(MLReader[CrossValidatorModel]):
|
|
cls: Type[CrossValidatorModel]
|
|
def __init__(self, cls: Type[CrossValidatorModel]) -> None: ...
|
|
def load(self, path: str) -> CrossValidatorModel: ...
|
|
|
|
class TrainValidationSplitWriter(MLWriter):
|
|
instance: TrainValidationSplit
|
|
def __init__(self, instance: TrainValidationSplit) -> None: ...
|
|
def saveImpl(self, path: str) -> None: ...
|
|
|
|
class TrainValidationSplitReader(MLReader[TrainValidationSplit]):
|
|
cls: Type[TrainValidationSplit]
|
|
def __init__(self, cls: Type[TrainValidationSplit]) -> None: ...
|
|
def load(self, path: str) -> TrainValidationSplit: ...
|
|
|
|
class TrainValidationSplitModelWriter(MLWriter):
|
|
instance: TrainValidationSplitModel
|
|
def __init__(self, instance: TrainValidationSplitModel) -> None: ...
|
|
def saveImpl(self, path: str) -> None: ...
|
|
|
|
class TrainValidationSplitModelReader(MLReader[TrainValidationSplitModel]):
|
|
cls: Type[TrainValidationSplitModel]
|
|
def __init__(self, cls: Type[TrainValidationSplitModel]) -> None: ...
|
|
def load(self, path: str) -> TrainValidationSplitModel: ...
|