def main():
sagemaker_session = sagemaker.Session()
stepfunctions.set_stream_logger(level=logging.INFO)
bucket = 's3://pixiv-image-backet'
sagemaker_execution_role = 'arn:aws:iam::829044821271:role/service-role/AmazonSageMaker-ExecutionRole-20200412T194702'
workflow_execution_role = 'arn:aws:iam::829044821271:role/StepFunctionsWorkflowExecutionRole'
estimator1 = PyTorch(entry_point='train.py',
source_dir='projection_discriminator',
role=sagemaker_execution_role,
framework_version='1.4.0',
train_instance_count=2,
train_instance_type='ml.m5.2xlarge',
hyperparameters={
'train_epoch': 1,
})
estimator2 = PyTorch(entry_point='train.py',
source_dir='wgan_gp',
role=sagemaker_execution_role,
framework_version='1.4.0',
train_instance_count=2,
train_instance_type='ml.m5.2xlarge',
hyperparameters={
'train_epoch': 1,
})
training_step1 = steps.TrainingStep(state_id='Train Step1',
estimator=estimator1,
data={
'training': bucket,
},
job_name='PD-Train-{0}'.format(
uuid.uuid4()))
training_step2 = steps.TrainingStep(state_id='Train Step2',
estimator=estimator2,
data={
'training': bucket,
},
job_name='PD-Train-{0}'.format(
uuid.uuid4()))
parallel_state = steps.Parallel(state_id='Parallel', )
parallel_state.add_branch(training_step1)
parallel_state.add_branch(training_step2)
workflow_definition = steps.Chain([parallel_state])
workflow = Workflow(
name='MyTraining-{0}'.format(uuid.uuid4()),
definition=workflow_definition,
role=workflow_execution_role,
)
workflow.create()
workflow.execute()
role=workflow_execution_role,
execution_input=execution_input
)
workflow.create()
# Documentation states the following:
# Updates an existing state machine by modifying its definition and/or role. Executions started immediately after calling this method may use the previous definition and role.
import time
time.sleep(60) # wait 60 secs to allow the update of the workflow to complete. The method is not syncronous!
# Finally, run the workflow!
execution = workflow.execute(
inputs={
'TrainingJobName': training_job_name, # Each Sagemaker Job requires a unique name,
'ModelName': model_name # Each Model requires a unique name,
}
)
# now let's create the cloudformation template parameters file ready for the CodeDeploy step in the pipeline
model_data_url = 's3://{}/{}/output'.format(bucket, project_name) + "/" + training_job_name + "/output/model.tar.gz"
parameter_file_data = {
"Parameters" : {
"ModelName" : model_name,
"ModelDataUrl" : model_data_url,
"TrainingImage": container,
"InstanceType" : "ml.t2.xlarge",
"InstanceCount": "1",
}
)
## SageMaker の学習ジョブを実行するステップ
estimator = create_estimator()
data_path = {'train': args.data_path}
training_step = steps.TrainingStep(
'Train Step',
estimator=estimator,
data=data_path,
job_name=execution_input['TrainJobName'],
wait_for_completion=False # SFnを実行した後に Bitbucket へプルリクを上げるように変更したため、ここは True で良いかも。
)
# 各 Step を連結
chain_list = [etl_step, training_step]
workflow_definition = steps.Chain(chain_list)
# Workflow の作成
workflow = Workflow(
name=FLOW_NAME,
definition=workflow_definition,
role=WORKFLOW_ROLE,
execution_input=execution_input
)
workflow.create()
# Workflow の実行
execution = workflow.execute(inputs=inputs)
name="Main-workflow",
definition=Main_workflow_definition,
role=workflow_execution_role
)
# COMMAND ----------
Main_workflow.render_graph()
# COMMAND ----------
Main_workflow.create()
# COMMAND ----------
Main_workflow_execution = Main_workflow.execute()
# COMMAND ----------
# MAGIC %md
# MAGIC Main_workflow_execution = Workflow(
# MAGIC name="Campaign_Workflow",
# MAGIC definition=path1,
# MAGIC role=workflow_execution_role
# MAGIC )
# COMMAND ----------
# COMMAND ----------