Vertex AI Experiments

start_execution.py

training_data_artifact = aiplatform.Artifact.create(
        schema_title='system.Dataset',
        uri='gs://doit-vertex-demo/higgs/training.csv',
        display_name='data')

with aiplatform.start_execution(
       schema_title="system.ContainerExecution", 
       display_name='training'
    ) as execution:

        execution.assign_input_artifacts([training_data_artifact])

        # do the training and store the model as output artifact

        model_artifact = aiplatform.Artifact.create(
        schema_title='system.Model',
        uri='gs://doit-vertex-demo/higgs/model',
        display_name='model')

        execution.assign_output_artifacts([model_artifact])