invoke-ai/InvokeAI · error · ValueError

There are no submodels in a TI model.

Error message

There are no submodels in a TI model.

What it means

A textual inversion (TI/embedding) model is a single artifact with no submodels, so the loader raises this ValueError whenever a submodel_type is provided. The message emphasizes that submodel requests are meaningless for TI embeddings.

Source

Thrown at invokeai/backend/model_manager/load/model_loaders/textual_inversion.py:33

    SubModelType,
)
from invokeai.backend.textual_inversion import TextualInversionModelRaw


@ModelLoaderRegistry.register(base=BaseModelType.Any, type=ModelType.TextualInversion, format=ModelFormat.EmbeddingFile)
@ModelLoaderRegistry.register(
    base=BaseModelType.Any, type=ModelType.TextualInversion, format=ModelFormat.EmbeddingFolder
)
class TextualInversionLoader(ModelLoader):
    """Class to load TI models."""

    def _load_model(
        self,
        config: AnyModelConfig,
        submodel_type: Optional[SubModelType] = None,
    ) -> AnyModel:
        if submodel_type is not None:
            raise ValueError("There are no submodels in a TI model.")
        model = TextualInversionModelRaw.from_checkpoint(
            file_path=config.path,
            dtype=self._torch_dtype,
        )
        return model

    # override
    def _get_model_path(self, config: AnyModelConfig) -> Path:
        model_path = self._app_config.models_path / config.path

        if config.format == ModelFormat.EmbeddingFolder:
            path = model_path / "learned_embeds.bin"
        else:
            path = model_path

        if not path.exists():
            raise OSError(f"The embedding file at {path} was not found")

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Pass submodel_type=None when loading TI embeddings.
2. Apply TI embeddings at pipeline level (prompt/concept loading), not via submodel requests.
3. Filter TI models out of any submodel-loading loops in your code.

Example fix

// before
model = loader.load_model(ti_config, submodel_type=SubModelType.TextEncoder)
// after
model = loader.load_model(ti_config, submodel_type=None)

if model_type is ModelType.TextualInversion and submodel_type is not None:
    submodel_type = None  # TI embeddings have no submodels

def is_embedding(model_type: ModelType) -> bool:
    return model_type is ModelType.TextualInversion

try:
    model = loader.load_model(ti_config, submodel_type=None)
except ValueError as e:
    logger.error("TI load failed: %s", e)
    raise

When it happens

Trigger: Calling load_model on a textual inversion config with submodel_type set (e.g. trying to fetch its 'text encoder' component) instead of None.

Common situations: Code iterating submodels of a main model that also touches TI embeddings; mistaking an embedding entry for a pipeline entry; automation that always requests a TextEncoder submodel.

Related errors

• There are no submodels in a LoRA model.
• Unexpected submodel requested for LLaVA OneVision model.
• Unexpected submodel requested for Spandrel model.
• Unexpected submodel requested for TextLLM model.
• Invalid or expired token

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/12812221abe7d8ae. Report an issue: GitHub.