invoke-ai/InvokeAI · error · ValueError

Only TextEncoder and Tokenizer submodels are supported. Rece

Error message

Only TextEncoder and Tokenizer submodels are supported. Received: {submodel_str}

What it means

For SDNQ Qwen3 encoder configs, _load_model only handles SubModelType.TextEncoder and SubModelType.Tokenizer. Requesting any other submodel (or None) from this loader raises a ValueError naming the received submodel type.

Source

Thrown at invokeai/backend/model_manager/load/model_loaders/z_image.py:1480

    def _load_model(
        self,
        config: AnyModelConfig,
        submodel_type: Optional[SubModelType] = None,
    ) -> AnyModel:
        if not isinstance(config, (Qwen3Encoder_SDNQ_Config, Qwen3Encoder_SDNQ_Folder_Config)):
            raise ValueError(
                "Only Qwen3Encoder_SDNQ_Config or Qwen3Encoder_SDNQ_Folder_Config models are supported here."
            )

        match submodel_type:
            case SubModelType.TextEncoder:
                return self._load_from_sdnq(config)
            case SubModelType.Tokenizer:
                return self._load_tokenizer_with_offline_fallback()

        submodel_str = submodel_type.value if submodel_type else "None"
        raise ValueError(f"Only TextEncoder and Tokenizer submodels are supported. Received: {submodel_str}")

    def _load_tokenizer_with_offline_fallback(self) -> AnyModel:
        """Load tokenizer with local_files_only fallback for offline support."""
        try:
            return AutoTokenizer.from_pretrained(self.DEFAULT_TOKENIZER_SOURCE, local_files_only=True)
        except OSError:
            return AutoTokenizer.from_pretrained(self.DEFAULT_TOKENIZER_SOURCE)

    def _load_from_sdnq(
        self,
        config: AnyModelConfig,
    ) -> AnyModel:
        from transformers import Qwen3Config, Qwen3ForCausalLM

        from invokeai.backend.quantization.sdnq.sdnq_tensor import SDNQTensor
        from invokeai.backend.util.logging import InvokeAILogger

        logger = InvokeAILogger.get_logger(self.__class__.__name__)

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Load only TextEncoder/Tokenizer submodels from this config; get VAE/transformer from their own model records.
2. Fix the submodel_type recorded for the model in the model manager DB (re-import the model).
3. Check the calling pipeline/graph code that resolves SubModelType values.
4. Guard caller code to skip unsupported submodel types for this model family.

Example fix

// before
model = loader._load_model(cfg, SubModelType.Vae)
// after
model = loader._load_model(cfg, SubModelType.TextEncoder)

SUPPORTED = {SubModelType.TextEncoder, SubModelType.Tokenizer}
if submodel_type not in SUPPORTED:
    raise ValueError(f"Z-Image SDNQ encoder supports only TextEncoder/Tokenizer, got {submodel_type}")

def supports_submodel(st: SubModelType | None) -> bool:
    return st in (SubModelType.TextEncoder, SubModelType.Tokenizer)

try:
    model = loader._load_model(cfg, submodel_type)
except ValueError as e:
    if "Only TextEncoder and Tokenizer" in str(e):
        model = load_from_other_model_record(submodel_type)  # VAE/transformer live elsewhere
    else:
        raise

When it happens

Trigger: Calling _load_model(config, submodel_type) with e.g. SubModelType.Vae, SubModelType.Main, or None for a Qwen3Encoder SDNQ config — typically a pipeline assembled the submodel list incorrectly or the model record's type is wrong.

Common situations: Model manager DB entry typed as the wrong SubModelType; custom workflow/graph node requesting the wrong submodel from this loader; API call loading submodels for a model that only exposes text encoder + tokenizer.

Related errors

• A submodel type must be provided when loading Ideogram 4 mai
• Unsupported submodel for Ideogram 4: {submodel_type.value if
• There are no submodels in an IP-Adapter model.
• Only Main_SDNQ_ZImage_Config or Main_SDNQ_Diffusers_ZImage_C
• Single-file SDNQ Z-Image checkpoints only provide the Transf

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/bd811a8074e22923. Report an issue: GitHub.