def testCloudTunerHyperparameters(self):
"""Test case to configure Tuner with HyperParameters object."""
study_id = "{}_hyperparameters".format(_STUDY_ID_BASE)
self._study_id = study_id
tuner = CloudTuner(
_build_model,
project_id=_PROJECT_ID,
region=_REGION,
objective="acc",
hyperparameters=_HPS,
max_trials=5,
study_id=study_id,
directory=os.path.join(self.get_temp_dir(), study_id),
)
# "Search space summary" comes first, but the order of
# "learning_rate (Float)" and "num_layers (Int)" is not deterministic,
# hence they are wrapped as look-ahead assertions in the regex.
self._assert_output(
tuner.search_space_summary,
r".*Search space summary(?=.*learning_rate \(Float\))"
r"(?=.*num_layers \(Int\).*)",
)
tuner.search(
x=self._x,
y=self._y,
epochs=2,
steps_per_epoch=20,
validation_steps=10,
validation_data=(self._val_x, self._val_y),
)
self._assert_results_summary(tuner.results_summary)
def _dist_search_fn(temp_dir, study_id, tuner_id):
"""Multi-process safe tuner instantiation and tuner.search()."""
# Jitter instantiation so as to avoid contention on study
# creation and dataset download.
time.sleep(int(tuner_id[5:])) # tuner_id is formatted as 'tuner%d'
# Dataset must be loaded independently in sub-process.
(x, y), (val_x, val_y) = _load_data(temp_dir)
tuner = CloudTuner(
_build_model,
project_id=_PROJECT_ID,
region=_REGION,
objective="acc",
hyperparameters=_HPS,
max_trials=5,
study_id=study_id,
directory=os.path.join(temp_dir, study_id, tuner_id),
)
tuner.tuner_id = tuner_id
tuner.search(
x=x,
y=y,
epochs=2,
steps_per_epoch=20,
validation_steps=10,
validation_data=(val_x, val_y),
)
return tuner
def tuner_fn(fn_args: FnArgs) -> TunerFnResult:
"""Build the tuner using the CloudTuner API.
Args:
fn_args: Holds args as name/value pairs. See
https://www.tensorflow.org/tfx/api_docs/python/tfx/components/trainer/fn_args_utils/FnArgs.
- transform_graph_path: optional transform graph produced by TFT.
- custom_config: An optional dictionary passed to the component. In this
example, it contains the dict ai_platform_tuning_args.
- working_dir: working dir for tuning.
- train_files: List of file paths containing training tf.Example data.
- eval_files: List of file paths containing eval tf.Example data.
- train_steps: number of train steps.
- eval_steps: number of eval steps.
Returns:
A namedtuple contains the following:
- tuner: A BaseTuner that will be used for tuning.
- fit_kwargs: Args to pass to tuner's run_trial function for fitting the
model , e.g., the training and validation dataset. Required
args depend on the above tuner's implementation.
"""
transform_graph = tft.TFTransformOutput(fn_args.transform_graph_path)
# CloudTuner is a subclass of kerastuner.Tuner which inherits from
# BaseTuner.
tuner = CloudTuner(
_build_keras_model,
# The project/region configuations for Cloud Vizier service and its trial
# executions. Note: this example uses the same configuration as the
# CAIP Training service for distributed tuning flock management to view
# all of the pipeline's jobs and resources in the same project. It can
# also be configured separately.
project_id=fn_args.custom_config['ai_platform_tuning_args']['project'],
region=fn_args.custom_config['ai_platform_tuning_args']['region'],
objective=kerastuner.Objective('val_sparse_categorical_accuracy',
'max'),
hyperparameters=_get_hyperparameters(),
max_trials=8, # Optional.
directory=fn_args.working_dir)
train_dataset = _input_fn(fn_args.train_files,
fn_args.data_accessor,
transform_graph,
batch_size=_TRAIN_BATCH_SIZE)
eval_dataset = _input_fn(fn_args.eval_files,
fn_args.data_accessor,
transform_graph,
batch_size=_EVAL_BATCH_SIZE)
return TunerFnResult(tuner=tuner,
fit_kwargs={
'x': train_dataset,
'validation_data': eval_dataset,
'steps_per_epoch': fn_args.train_steps,
'validation_steps': fn_args.eval_steps
})
def testCloudTunerStudyConfig(self):
"""Test case to configure Tuner with StudyConfig object."""
# Configure the search space. Specification:
# https://cloud.google.com/ai-platform/optimizer/docs/reference/rest/v1/projects.locations.studies#StudyConfig # pylint: disable=line-too-long
study_config = {
"metrics": [{"goal": "MAXIMIZE", "metric": "acc"}],
"parameters": [
{
"discrete_value_spec": {"values": [0.0001, 0.001, 0.01]},
"parameter": "learning_rate",
"type": "DISCRETE",
},
{
"integer_value_spec": {"max_value": 10, "min_value": 2},
"parameter": "num_layers",
"type": "INTEGER",
},
{
"discrete_value_spec": {"values": [32, 64, 96, 128]},
"parameter": "units",
"type": "DISCRETE",
},
],
"algorithm": "ALGORITHM_UNSPECIFIED",
"automatedStoppingConfig": {
"decayCurveStoppingConfig": {"useElapsedTime": True}
},
}
study_id = "{}_study_config".format(_STUDY_ID_BASE)
self._study_id = study_id
tuner = CloudTuner(
_build_model,
project_id=_PROJECT_ID,
region=_REGION,
study_config=study_config,
study_id=study_id,
max_trials=5,
directory=os.path.join(self.get_temp_dir(), study_id),
)
self._assert_output(
tuner.search_space_summary,
r".*Search space summary(?=.*learning_rate \(Choice\))"
r"(?=.*num_layers \(Int\))(?=.*units \(Int\))",
)
tuner.search(
x=self._x,
y=self._y,
epochs=2,
steps_per_epoch=20,
validation_steps=10,
validation_data=(self._val_x, self._val_y),
)
self._assert_results_summary(tuner.results_summary)
def tuner_fn(fn_args: TrainerFnArgs) -> TunerFnResult:
"""Build the tuner using CloudTuner (KerasTuner instance).
Args:
fn_args: Holds args as name/value pairs.
- working_dir: working dir for tuning.
- train_files: List of file paths containing training tf.Example data.
- eval_files: List of file paths containing eval tf.Example data.
- train_steps: number of train steps.
- eval_steps: number of eval steps.
- schema_path: optional schema of the input data.
- transform_graph_path: optional transform graph produced by TFT.
Returns:
A namedtuple contains the following:
- tuner: A BaseTuner that will be used for tuning.
- fit_kwargs: Args to pass to tuner's run_trial function for fitting the
model , e.g., the training and validation dataset. Required
args depend on the above tuner's implementation.
"""
transform_graph = tft.TFTransformOutput(fn_args.transform_graph_path)
# Construct a build_keras_model_fn that just takes hyperparams from get_hyperparameters as input.
build_keras_model_fn = functools.partial(
_build_keras_model, tf_transform_output=transform_graph)
# CloudTuner is a subclass of kerastuner.Tuner which inherits from BaseTuner.
tuner = CloudTuner(build_keras_model_fn,
project_id=Config.PROJECT_ID,
region=Config.GCP_REGION,
max_trials=50,
hyperparameters=_get_hyperparameters(),
objective=kerastuner.Objective(
'val_sparse_categorical_accuracy', 'max'),
directory=fn_args.working_dir)
train_dataset = _input_fn(fn_args.train_files,
fn_args.data_accessor,
transform_graph,
batch_size=TRAIN_BATCH_SIZE)
eval_dataset = _input_fn(fn_args.eval_files,
fn_args.data_accessor,
transform_graph,
batch_size=EVAL_BATCH_SIZE)
return TunerFnResult(tuner=tuner,
fit_kwargs={
'x': train_dataset,
'validation_data': eval_dataset,
'steps_per_epoch': fn_args.train_steps,
'validation_steps': fn_args.eval_steps
})
def tuner_fn(fn_args: TrainerFnArgs) -> TunerFnResult:
"""Build the tuner using CloudTuner (KerasTuner instance).
Args:
fn_args: Holds args used to train and tune the model as name/value pairs. See
https://www.tensorflow.org/tfx/api_docs/python/tfx/components/trainer/fn_args_utils/FnArgs.
Returns:
A namedtuple contains the following:
- tuner: A BaseTuner that will be used for tuning.
- fit_kwargs: Args to pass to tuner's run_trial function for fitting the
model , e.g., the training and validation dataset. Required
args depend on the above tuner's implementation.
"""
transform_graph = tft.TFTransformOutput(fn_args.transform_graph_path)
# Construct a build_keras_model_fn that just takes hyperparams from get_hyperparameters as input.
build_keras_model_fn = functools.partial(
_build_keras_model, tf_transform_output=transform_graph)
# CloudTuner is a subclass of kerastuner.Tuner which inherits from BaseTuner.
tuner = CloudTuner(
build_keras_model_fn,
project_id=fn_args.custom_config['ai_platform_training_args']['project'],
region=fn_args.custom_config['ai_platform_training_args']['region'],
max_trials=50,
hyperparameters=_get_hyperparameters(),
objective=kerastuner.Objective('val_sparse_categorical_accuracy', 'max'),
directory=fn_args.working_dir)
train_dataset = _input_fn(
fn_args.train_files,
fn_args.data_accessor,
transform_graph,
batch_size=TRAIN_BATCH_SIZE)
eval_dataset = _input_fn(
fn_args.eval_files,
fn_args.data_accessor,
transform_graph,
batch_size=EVAL_BATCH_SIZE)
return TunerFnResult(
tuner=tuner,
fit_kwargs={
'x': train_dataset,
'validation_data': eval_dataset,
'steps_per_epoch': fn_args.train_steps,
'validation_steps': fn_args.eval_steps
})
def testCloudTunerDatasets(self):
"""Test case to configure Tuner with tf.data.Dataset as input data."""
train_dataset = (
tf.data.Dataset.from_tensor_slices((self._x, self._y))
.batch(128)
.cache()
.prefetch(1000)
)
eval_dataset = (
tf.data.Dataset.from_tensor_slices((self._val_x, self._val_y))
.batch(128)
.cache()
.prefetch(1000)
)
study_id = "{}_dataset".format(_STUDY_ID_BASE)
self._study_id = study_id
tuner = CloudTuner(
_build_model,
project_id=_PROJECT_ID,
region=_REGION,
objective="acc",
hyperparameters=_HPS,
study_id=study_id,
max_trials=5,
directory=os.path.join(self.get_temp_dir(), study_id),
)
self._assert_output(
tuner.search_space_summary,
r".*Search space summary(?=.*learning_rate \(Float\))"
r"(?=.*num_layers \(Int\).*)",
)
tuner.search(
x=train_dataset,
epochs=2,
steps_per_epoch=20,
validation_steps=10,
validation_data=eval_dataset,
)
self._assert_results_summary(tuner.results_summary)
def tuner_fn(fn_args: FnArgs) -> TunerFnResult:
"""Build the tuner using the CloudTuner API.
Args:
fn_args: Holds args as name/value pairs.
- working_dir: working dir for tuning.
- train_files: List of file paths containing training tf.Example data.
- eval_files: List of file paths containing eval tf.Example data.
- train_steps: number of train steps.
- eval_steps: number of eval steps.
- schema_path: optional schema of the input data.
- transform_graph_path: optional transform graph produced by TFT.
Returns:
A namedtuple contains the following:
- tuner: A BaseTuner that will be used for tuning.
- fit_kwargs: Args to pass to tuner's run_trial function for fitting the
model , e.g., the training and validation dataset. Required
args depend on the above tuner's implementation.
"""
# CloudTuner is a subclass of kerastuner.Tuner which inherits from
# BaseTuner.
tuner = CloudTuner(
_build_keras_model,
project_id=_PROJECT_ID,
region=_REGION,
objective=kerastuner.Objective('val_sparse_categorical_accuracy', 'max'),
hyperparameters=_get_hyperparameters(),
max_trials=8, # Optional.
directory=fn_args.working_dir
)
transform_graph = tft.TFTransformOutput(fn_args.transform_graph_path)
train_dataset = _input_fn(fn_args.train_files, transform_graph)
eval_dataset = _input_fn(fn_args.eval_files, transform_graph)
return TunerFnResult(
tuner=tuner,
fit_kwargs={
'x': train_dataset,
'validation_data': eval_dataset,
'steps_per_epoch': fn_args.train_steps,
'validation_steps': fn_args.eval_steps
})
def tuner_fn(fn_args: FnArgs) -> TunerFnResult:
"""Build the tuner using CloudTuner (KerasTuner instance).
Args:
fn_args: Holds args used to train and tune the model as name/value pairs. See
https://www.tensorflow.org/tfx/api_docs/python/tfx/components/trainer/fn_args_utils/FnArgs.
Returns:
A namedtuple contains the following:
- tuner: A BaseTuner that will be used for tuning.
- fit_kwargs: Args to pass to tuner's run_trial function for fitting the
model , e.g., the training and validation dataset. Required
args depend on the above tuner's implementation.
"""
custom_config_dict = _get_custom_config_dict(fn_args)
max_trials = custom_config_dict.get('max_trials', MAX_TRIALS)
transform_graph = tft.TFTransformOutput(fn_args.transform_graph_path)
# Construct a build_keras_model_fn that just takes hyperparams from get_hyperparameters as input.
build_keras_model_fn = functools.partial(
_build_keras_model, tf_transform_output=transform_graph)
# CloudTuner is a subclass of kerastuner.Tuner which inherits from BaseTuner.
#is_local_run = "custom_config" not in fn_args.custom_config
is_local_run = custom_config_dict.get("is_local_run", True)
absl.logging.info('is_local_run : %s' % is_local_run)
if is_local_run:
tuner = kerastuner.RandomSearch(build_keras_model_fn,
max_trials=max_trials,
hyperparameters=_get_hyperparameters(),
allow_new_entries=False,
objective=kerastuner.Objective(
'val_binary_accuracy', 'max'),
directory=fn_args.working_dir,
project_name='titanic_tuning')
else:
tuner = CloudTuner(
build_keras_model_fn,
project_id=fn_args.custom_config['ai_platform_training_args']
['project'],
region=fn_args.custom_config['ai_platform_training_args']
['region'],
max_trials=max_trials,
hyperparameters=_get_hyperparameters(),
objective=kerastuner.Objective('val_binary_accuracy', 'max'),
# objective=kerastuner.Objective('auc', 'min'),
directory=fn_args.working_dir)
train_dataset = _input_fn(fn_args.train_files,
fn_args.data_accessor,
transform_graph,
batch_size=TRAIN_BATCH_SIZE)
eval_dataset = _input_fn(fn_args.eval_files,
fn_args.data_accessor,
transform_graph,
batch_size=EVAL_BATCH_SIZE)
return TunerFnResult(tuner=tuner,
fit_kwargs={
'x': train_dataset,
'validation_data': eval_dataset,
'steps_per_epoch': fn_args.train_steps,
'validation_steps': fn_args.eval_steps
})