invoke-ai/InvokeAI · error · ValueError

Only CheckpointConfigBase models are currently supported her

Error message

Only CheckpointConfigBase models are currently supported here.

What it means

The Qwen Image single-file (checkpoint) loader only accepts configs deriving from CheckpointConfigBase. When _load_model receives a diffusers-folder style config (e.g. Main_Diffusers_Config or a GGUF config class not based on checkpoint), it raises this ValueError because the single-file loading path cannot handle it.

Source

Thrown at invokeai/backend/model_manager/load/model_loaders/qwen_image.py:192

                result = load_class.from_pretrained(model_path, **dtype_kwarg, local_files_only=True)
            else:
                raise e

        result = self._apply_fp8_layerwise_casting(result, config, submodel_type)
        return result


@ModelLoaderRegistry.register(base=BaseModelType.QwenImage, type=ModelType.Main, format=ModelFormat.GGUFQuantized)
class QwenImageGGUFCheckpointModel(ModelLoader):
    """Class to load GGUF-quantized Qwen Image Edit transformer models."""

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

        match submodel_type:
            case SubModelType.Transformer:
                return self._load_from_singlefile(config)

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

    def _load_from_singlefile(self, config: AnyModelConfig) -> AnyModel:
        from diffusers import QwenImageTransformer2DModel

        if not isinstance(config, Main_GGUF_QwenImage_Config):
            raise TypeError(f"Expected Main_GGUF_QwenImage_Config, got {type(config).__name__}.")
        model_path = Path(config.path)

        target_device = TorchDevice.choose_torch_device()
        compute_dtype = TorchDevice.choose_bfloat16_safe_dtype(target_device)

View on GitHub (pinned to 0b6a024f2f)

Solutions

Re-import or re-convert the model so it is registered with a CheckpointConfigBase-derived config (Main_Checkpoint_QwenImage_Config or Main_GGUF_QwenImage_Config).
Verify the model's format in the model manager UI/API matches its actual file layout (single-file vs diffusers folder).
If the model is truly a diffusers folder, use the diffusers-registered Qwen Image loader instead.

Example fix

// before
config = Main_Diffusers_Config(path=..., model_format=...)  # folder-style config
model = checkpoint_loader._load_model(config, SubModelType.Transformer)
// after
config = Main_Checkpoint_QwenImage_Config(path=model_file, config=ConfigVariantEnum.QwenImage)
model = checkpoint_loader._load_model(config, SubModelType.Transformer)

Defensive patterns

Strategy: type-guard

Validate before calling

from invokeai.backend.model_manager.config import Checkpoint_Config_Base
if not isinstance(config, Checkpoint_Config_Base):
    raise TypeError(f"Single-file loader needs a checkpoint config, got {type(config).__name__}")

Type guard

def is_checkpoint_config(config: AnyModelConfig) -> bool:
    return isinstance(config, Checkpoint_Config_Base)

Try / catch

try:
    model = loader._load_model(config, SubModelType.Transformer)
except ValueError as e:
    if "Only CheckpointConfigBase" in str(e):
        config = reimport_model_with_correct_format(model_path)  # re-scan/re-register
        model = loader._load_model(config, SubModelType.Transformer)
    else:
        raise

Prevention

Match the model record's format field to the on-disk layout (single-file vs folder)
Re-scan the models directory after moving or converting model files
Check isinstance before calling format-specific loaders directly

When it happens

Trigger: Registering a Qwen Image single-file model but having it resolved with a non-checkpoint config class, or calling the checkpoint-registered loader's _load_model with a Diffusers/GGUF config object.

Common situations: Model format misdetected during scan (single-file .safetensors registered as diffusers); wrong ModelFormat on import; hand-constructed config passed to the wrong loader.

Related errors

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/1cb1dbd59db695db. Report an issue: GitHub.