Hi! Im trying pipelines with data creating and model training.
pipe = PipelineController(
name="data_preprocess_pipeline", project="some", version="1.0.0"
)
pipe.set_default_execution_queue("antifraude")
pipe.add_step(
name="data_creation",
base_task_project="some",
base_task_name="data_creation",
)
pipe.add_step(
name="yolov5_training",
base_task_project="some",
base_task_name="yolov5_jetson",
parents=["data_creation", ],
parameter_override={
"Args/dataset_id": "${data_creation.parameters.dataset_id}"}
)
pipe.start(queue="antifraude")

In data creating task i’ve done this: task.set_parameter("Args/dataset_id", dataset.id)
And in model training task this:
dataset_id = task.get_parameter("Args/dataset_id")

if dataset_id is not None:
dataset_path = f"clearml://{dataset_id}"
config.data = dataset_path

But ive got this error: ValueError: Could not locate parameter parameters.dataset_id on previous step data_creation

Whats wrong?