def testVertexDistributedTunerPipeline(self):
"""Tuner-only pipeline for distributed Tuner flock on Vertex AI Training."""
pipeline_name = self._make_unique_pipeline_name(
'kubeflow-vertex-dist-tuner')
pipeline = self._create_pipeline(
pipeline_name,
[
self.penguin_examples_importer,
self.penguin_schema_importer,
ai_platform_tuner_component.Tuner(
examples=self.penguin_examples_importer.outputs['result'],
module_file=self._penguin_tuner_module,
schema=self.penguin_schema_importer.outputs['result'],
train_args=trainer_pb2.TrainArgs(num_steps=10),
eval_args=trainer_pb2.EvalArgs(num_steps=5),
# 3 worker parallel tuning.
tune_args=tuner_pb2.TuneArgs(num_parallel_trials=3),
custom_config={
ai_platform_tuner_executor.TUNING_ARGS_KEY:
self._getVertexTrainingArgs(pipeline_name),
constants.ENABLE_VERTEX_KEY:
True,
constants.VERTEX_REGION_KEY:
self._GCP_REGION
})
])
self._compile_and_run_pipeline(pipeline)
self._assertHyperparametersAreWritten(pipeline_name)
def testConstructWithoutCustomConfig(self):
tuner = component.Tuner(
examples=self.examples,
schema=self.schema,
train_args=self.train_args,
eval_args=self.eval_args,
tune_args=self.tune_args,
module_file='/path/to/module/file',
)
self._verify_output(tuner)
def testAIPlatformDistributedTunerPipeline(self):
"""Tuner-only pipeline for distributed Tuner flock on AIP Training."""
pipeline_name = 'kubeflow-aip-dist-tuner-test-{}'.format(
test_utils.random_id())
pipeline = self._create_pipeline(
pipeline_name,
[
self.iris_examples_importer,
self.iris_schema_importer,
ai_platform_tuner_component.Tuner(
examples=self.iris_examples_importer.outputs['result'],
module_file=self._iris_tuner_module,
schema=self.iris_schema_importer.outputs['result'],
train_args=trainer_pb2.TrainArgs(num_steps=10),
eval_args=trainer_pb2.EvalArgs(num_steps=5),
# 3 worker parallel tuning.
tune_args=tuner_pb2.TuneArgs(num_parallel_trials=3),
custom_config={
ai_platform_trainer_executor.TRAINING_ARGS_KEY:
self._getCaipTrainingArgs(pipeline_name)
})
])
self._compile_and_run_pipeline(pipeline)
self._assertHyperparametersAreWritten(pipeline_name)