invoke-ai/InvokeAI · error · ValueError

Model '{main_config.name}' is not a Krea-2 main model. Selec

Error message

Model '{main_config.name}' is not a Krea-2 main model. Select a Krea-2 transformer model.

What it means

Krea2ModelLoader.invoke() requires the selected main model to have base model type Krea2 and type Main. Any other selection (wrong base family or non-main submodel) is rejected with this ValueError because the loader derives the Krea-2 transformer from the main model config.

Source

Thrown at invokeai/app/invocations/krea2_model_loader.py:78

        input=Input.Direct,
        ui_model_base=[BaseModelType.QwenImage, BaseModelType.Anima],
        ui_model_type=ModelType.VAE,
        title="VAE",
    )

    qwen3_vl_encoder_model: Optional[ModelIdentifierField] = InputField(
        default=None,
        description="Standalone Qwen3-VL Encoder model. "
        "If not provided, the encoder is loaded from the Krea-2 (diffusers) model.",
        input=Input.Direct,
        ui_model_type=ModelType.Qwen3VLEncoder,
        title="Qwen3-VL Encoder",
    )

    def invoke(self, context: InvocationContext) -> Krea2ModelLoaderOutput:
        main_config = context.models.get_config(self.model)
        if main_config.base is not BaseModelType.Krea2 or main_config.type is not ModelType.Main:
            raise ValueError(
                f"Model '{main_config.name}' is not a Krea-2 main model. Select a Krea-2 transformer model."
            )

        # Transformer always comes from the main model.
        transformer = self.model.model_copy(update={"submodel_type": SubModelType.Transformer})

        # Determine VAE source.
        if self.vae_model is not None:
            vae_config = context.models.get_config(self.vae_model)
            if vae_config.type is not ModelType.VAE or vae_config.base not in (
                BaseModelType.QwenImage,
                BaseModelType.Anima,
            ):
                raise ValueError(
                    f"VAE '{vae_config.name}' is not compatible with Krea-2. Select a Qwen Image or Anima VAE."
                )
            vae = self.vae_model.model_copy(update={"submodel_type": SubModelType.VAE})
        else:

View on GitHub (pinned to 0b6a024f2f)

Solutions

In the model manager, install/select a main model whose base is Krea2 and type is Main, then reference that in the loader node.
Check the model record: convert or re-import the checkpoint with base=Krea2 if it was misclassified.
If you meant a different architecture, use the matching loader invocation (e.g. Qwen Image or Flux loader) instead of Krea2ModelLoader.

Example fix

// before
loader = Krea2ModelLoader(model=qwen_image_main_model, ...)
// after
loader = Krea2ModelLoader(model=krea2_main_model, ...)  # base=Krea2, type=Main

Defensive patterns

Strategy: validation

Validate before calling

cfg = context.models.get_config(model_field)
if cfg.base is not BaseModelType.Krea2 or cfg.type is not ModelType.Main:
    raise ValueError(f"{cfg.name} is not a Krea-2 main model")

Type guard

from invokeai.backend.model_manager.config import BaseModelType, ModelType

def is_krea2_main(cfg) -> bool:
    return cfg.base is BaseModelType.Krea2 and cfg.type is ModelType.Main

Try / catch

try:
    output = loader.invoke(context)
except ValueError as e:
    if "not a Krea-2 main model" in str(e):
        output = pick_default_krea2_main(context).and_reload()
    else:
        raise

Prevention

Reference only top-level main model records in loader nodes, never submodel identifiers
Filter node model dropdowns by base=Krea2, type=Main
Re-verify model classifications after bulk imports or migrations

When it happens

Trigger: Passing a ModelIdentifierField to Krea2ModelLoader.model whose config has base != BaseModelType.Krea2, or type != ModelType.Main (e.g. a Qwen Image, SDXL, or Flux main model, or a VAE/encoder submodel reference).

Common situations: Selecting a model installed under a different base family in the model manager; pointing the loader at a submodel (transformer/VAE) record instead of the main model; stale model lists after re-install or base-type conversion.

Related errors

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/0816fd70301b8af4. Report an issue: GitHub.