HumanSignal/label-studio · error · ValueError

If you use "predictions" field in the task, you must put "da

Error message

If you use "predictions" field in the task, you must put "data" field in the task too

What it means

add_task enforces that tasks carrying pre-made predictions must also include the actual task payload. If a dict has a non-empty 'predictions' field but no 'data' key, the annotations/predictions cannot be attached to any data, so a ValueError is raised before the task is created.

Source

Thrown at label_studio/io_storages/base_models.py:502

        # | "AddObjects" >> label_studio_semantic_search.indexer.add_objects_from_bucket
        # --> add objects from batch to Vector DB
        # or for project task creation last step would be
        # | "AddObject" >> ImportStorage.add_task

        raise NotImplementedError

    @classmethod
    def add_task(cls, project, maximum_annotations, max_inner_id, storage, link_object: StorageObject, link_class):
        link_kwargs = asdict(link_object)
        data = link_kwargs.pop('task_data', None)

        allow_skip = data.get('allow_skip', True)

        # predictions
        predictions = data.get('predictions') or []
        if predictions:
            if 'data' not in data:
                raise ValueError(
                    'If you use "predictions" field in the task, you must put "data" field in the task too'
                )

        # annotations
        annotations = data.get('annotations') or []
        cancelled_annotations = 0
        if annotations:
            if 'data' not in data:
                raise ValueError(
                    'If you use "annotations" field in the task, you must put "data" field in the task too'
                )
            cancelled_annotations = len([a for a in annotations if a.get('was_cancelled', False)])

        storage_task_data_validator = load_func(getattr(settings, 'STORAGE_TASK_DATA_VALIDATOR', None))
        if storage_task_data_validator:
            storage_task_data_validator(project, data)

        if 'data' in data and isinstance(data['data'], dict):

View on GitHub (pinned to 0b49e9b539)

Solutions

1. Add the 'data' field (the task payload matching your labeling config) to every dict that includes 'predictions'.
2. Strip the 'predictions' field if you only want raw data imported and predictions are not needed.
3. Pre-validate your import JSON for the presence of 'data' whenever 'predictions' is present.

Example fix

// before
task = {'predictions': [{'result': [...], 'score': 0.9}]}
storage.add_task(task)

// after
task = {'data': {'image': 's3://bucket/img.jpg'}, 'predictions': [{'result': [...], 'score': 0.9}]}
storage.add_task(task)

def validate_task_payload(task: dict):
    if isinstance(task, dict) and (task.get('predictions') or task.get('annotations')) and 'data' not in task:
        raise ValueError('Task payload with predictions/annotations must also include "data"')

def is_valid_task_payload(task: dict) -> bool:
    return not (isinstance(task, dict) and (task.get('predictions') or task.get('annotations')) and 'data' not in task)

try:
    storage.add_task(task)
except ValueError as e:
    if '"data" field in the task' in str(e):
        logger.error('Malformed task payload (missing data): %s', task)
    else:
        raise

When it happens

Trigger: Calling add_task (directly or via storage sync _scan_and_create_links / create_tasks) with a dict like {'predictions': [...]} that omits 'data'.

Common situations: Importing JSON files exported from another project where data was stripped; hand-written imports that include only predictions; script-generated tasks with incomplete payloads.

Understand the failure class

Background: Schema validation failed / invalid input schema: payload rejected because its shape doesn't match the expected schema — this error's family across 28 libraries.

Related errors

• Prediction validation failed ({len(validation_errors)} error
• If you use "annotations" field in the task, you must put "da
• {'predictions': prediction_errors}
• "url" must be 2048 characters or fewer
• {item} contains invalid "task" field: task ID {task_id} not

AI-assisted analysis of HumanSignal/label-studio@0b49e9b539 (2026-08-29). Data as JSON: /api/errors/6e8c12bf800f3894. Report an issue: GitHub.