unslothai/unsloth · error · RecipeDatasetPublishError

NeMo Data Designer Hugging Face integration is not installed

Error message

NeMo Data Designer Hugging Face integration is not installed.

What it means

RecipeDatasetPublishError raised when importing data_designer.integrations.huggingface.* (HuggingFaceHubClient, dataset card classes, constants) fails with ImportError. The publish feature depends on the NeMo Data Designer Hugging Face integration being installed in the backend environment; without those modules, publishing cannot proceed and the original ImportError is chained as the cause.

Source

Thrown at studio/backend/core/data_recipe/huggingface.py:69

) -> str:
    dataset_path = _resolve_recipe_artifact_path(artifact_path)

    try:
        from data_designer.engine.storage.artifact_storage import (
            FINAL_DATASET_FOLDER_NAME,
            METADATA_FILENAME,
            PROCESSORS_OUTPUTS_FOLDER_NAME,
            SDG_CONFIG_FILENAME,
        )
        from data_designer.integrations.huggingface.client import (
            HuggingFaceHubClient,
            HuggingFaceHubClientUploadError,
        )
        from data_designer.integrations.huggingface.dataset_card import (
            DataDesignerDatasetCard,
        )
    except ImportError as exc:
        raise RecipeDatasetPublishError(
            "NeMo Data Designer Hugging Face integration is not installed."
        ) from exc

    try:
        client = HuggingFaceHubClient(token = hf_token)
        client._validate_repo_id(repo_id = repo_id)
        client._validate_dataset_path(base_dataset_path = dataset_path)
        client._create_or_get_repo(repo_id = repo_id, private = private)

        metadata_path = dataset_path / METADATA_FILENAME
        builder_config_path = dataset_path / SDG_CONFIG_FILENAME

        with metadata_path.open(encoding = "utf-8") as fh:
            metadata = json.load(fh)

        builder_config = None
        if builder_config_path.exists():
            with builder_config_path.open(encoding = "utf-8") as fh:

View on GitHub (pinned to 203007d190)

Solutions

  1. Install the extra that provides data_designer.integrations.huggingface (e.g. the NeMo Data Designer / nemo-toolkit package with its HF extras) into the backend environment.
  2. Pin a version of the dependency known to contain integrations.huggingface.client.HuggingFaceHubClient.
  3. Verify with: python -c "from data_designer.integrations.huggingface.client import HuggingFaceHubClient".
  4. If you do not need HF publishing, avoid the publish endpoint instead of installing.

Example fix

# before: backend env missing the integration
pip install nemo-toolkit  # no HF integration extras

# after
pip install 'nemo-toolkit[data-designer]'  # or the extra that ships data_designer.integrations.huggingface
python -c "from data_designer.integrations.huggingface.client import HuggingFaceHubClient"
Defensive patterns

Strategy: validation

Validate before calling

try:
    from data_designer.integrations.huggingface.client import HuggingFaceHubClient  # noqa: F401
    hf_integration_ok = True
except ImportError:
    hf_integration_ok = False

Try / catch

try:
    publish_recipe_dataset(...)
except RecipeDatasetPublishError as e:
    if 'not installed' in str(e):
        disable_publish_ui(); show_install_instructions()

Prevention

When it happens

Trigger: Calling publish_recipe_dataset in an environment where the 'data_designer' package (NeMo Data Designer / NeMo-Toolkit extras) is missing or lacks the integrations.huggingface submodule; a version of the package that moved/renamed those modules; a broken install with partially installed dependencies.

Common situations: Custom deployment images that trimmed optional dependencies; upgrading NeMo-Toolkit to a version where the integration moved; running the backend from a different venv than the one with the extras installed.

Related errors


AI-assisted analysis of unslothai/unsloth@203007d190 (2026-08-15). Data as JSON: /api/errors/1dfbb96f2185132f. Report an issue: GitHub.