invoke-ai/InvokeAI · error · TypeError

Expected QwenVLEncoder_Checkpoint_Config, got {type(config).

Error message

Expected QwenVLEncoder_Checkpoint_Config, got {type(config).__name__}.

What it means

The checkpoint-format Qwen VL encoder loader only accepts QwenVLEncoder_Checkpoint_Config. Any other config type reaching its _load_model raises this TypeError naming the actual type, because its helpers load tokenizer/text-encoder weights from a single checkpoint file rather than a diffusers folder.

Source

Thrown at invokeai/backend/model_manager/load/model_loaders/qwen_image.py:347

@ModelLoaderRegistry.register(base=BaseModelType.Any, type=ModelType.QwenVLEncoder, format=ModelFormat.Checkpoint)
class QwenVLEncoderCheckpointLoader(ModelLoader):
    """Loads a single-file Qwen2.5-VL encoder checkpoint (e.g. ComfyUI fp8_scaled).

    The checkpoint bundles the language model and the visual tower into one
    safetensors file. Tokenizer + processor are pulled from HuggingFace
    (`Qwen/Qwen2.5-VL-7B-Instruct`) on first use, with offline cache fallback.
    """

    DEFAULT_HF_REPO = "Qwen/Qwen2.5-VL-7B-Instruct"

    def _load_model(
        self,
        config: AnyModelConfig,
        submodel_type: Optional[SubModelType] = None,
    ) -> AnyModel:
        if not isinstance(config, QwenVLEncoder_Checkpoint_Config):
            raise TypeError(f"Expected QwenVLEncoder_Checkpoint_Config, got {type(config).__name__}.")

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

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

    def _load_tokenizer_with_offline_fallback(self) -> AnyModel:
        from transformers import AutoTokenizer

        from invokeai.backend.util.logging import InvokeAILogger

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

View on GitHub (pinned to 0b6a024f2f)

Solutions

Use QwenVLEncoder_Checkpoint_Config when calling this loader, with path pointing at the single checkpoint file.
If the model is a diffusers folder, route it to the diffusers-registered QwenVLEncoder loader.
Re-scan or edit the model record so format and config class agree with the on-disk layout.

Example fix

// before
config = QwenVLEncoder_Diffusers_Config(path="encoder_dir/")
enc = checkpoint_loader._load_model(config, SubModelType.TextEncoder)
// after
config = QwenVLEncoder_Checkpoint_Config(path="encoder.safetensors")
enc = checkpoint_loader._load_model(config, SubModelType.TextEncoder)

Defensive patterns

Strategy: type-guard

Validate before calling

from invokeai.backend.model_manager.config import QwenVLEncoder_Checkpoint_Config
if not isinstance(config, QwenVLEncoder_Checkpoint_Config):
    raise TypeError(f"Checkpoint QwenVL loader needs QwenVLEncoder_Checkpoint_Config, got {type(config).__name__}")

Type guard

def is_checkpoint_qwenvl_config(config: AnyModelConfig) -> bool:
    return isinstance(config, QwenVLEncoder_Checkpoint_Config)

Try / catch

try:
    enc = loader._load_model(config, SubModelType.TextEncoder)
except TypeError as e:
    if "QwenVLEncoder_Checkpoint_Config" in str(e):
        enc = diffusers_qwenvl_loader._load_model(config, SubModelType.TextEncoder)
    else:
        raise

Prevention

Ensure checkpoint-format encoders use QwenVLEncoder_Checkpoint_Config
Point the config path at the single checkpoint file, not a folder
Keep format and config class consistent in model records

When it happens

Trigger: Passing a QwenVLEncoder_Diffusers_Config (or any other config) into the checkpoint-registered loader's _load_model, e.g. via a model record whose format says checkpoint but whose config is diffusers-derived.

Common situations: Model record format/config mismatch after import; diffusers-folder encoder incorrectly registered as checkpoint; scripts constructing the wrong config class.

Related errors

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/d6e5ff4d294eb90b. Report an issue: GitHub.