invoke-ai/InvokeAI · error · ValueError

An expected config.json file is missing from this model.

Error message

An expected config.json file is missing from this model.

What it means

When get_hf_load_class resolves the diffusers/transformers class, it looks up the class name in the library's module maps. A KeyError means the referenced class or expected config entry does not exist; this except clause converts it to a ValueError stating that the model's config.json (or a key within it expected by the maps) is missing from the model. It signals the model directory is incomplete relative to what the loader expects.

Source

Thrown at invokeai/backend/model_manager/load/model_loaders/generic_diffusers.py:74

        result = None
        if submodel_type:
            try:
                config = self._load_diffusers_config(model_path, config_name="model_index.json")
                module, class_name = config[submodel_type.value]
                result = self._hf_definition_to_type(module=module, class_name=class_name)
            except KeyError as e:
                raise ValueError(f'The "{submodel_type}" submodel is not available for this model.') from e
        else:
            try:
                config = self._load_diffusers_config(model_path, config_name="config.json")
                if class_name := config.get("_class_name"):
                    result = self._hf_definition_to_type(module="diffusers", class_name=class_name)
                elif class_name := config.get("architectures"):
                    result = self._hf_definition_to_type(module="transformers", class_name=class_name[0])
                else:
                    raise RuntimeError("Unable to decipher Load Class based on given config.json")
            except KeyError as e:
                raise ValueError("An expected config.json file is missing from this model.") from e
        assert result is not None
        return result

    # TO DO: Add exception handling
    def _hf_definition_to_type(self, module: str, class_name: str) -> ModelMixin:  # fix with correct type
        if module in [
            "diffusers",
            "transformers",
            "invokeai.backend.quantization.fast_quantized_transformers_model",
            "invokeai.backend.quantization.fast_quantized_diffusion_model",
        ]:
            res_type = sys.modules[module]
        else:
            res_type = sys.modules["diffusers"].pipelines
        result: ModelMixin = getattr(res_type, class_name)
        return result

    def _load_diffusers_config(self, model_path: Path, config_name: str = "config.json") -> dict[str, Any]:

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Check the model directory actually contains config.json and re-download it if missing.
2. Add a complete diffusers-style config.json (with `_class_name`) to the model directory.
3. If the class name in config is non-standard, rename it to the canonical diffusers/transformers class name.
4. Re-export the model using the official diffusers conversion script so all metadata files are produced.

Example fix

// before: model dir contains only model.safetensors
// after: ensure model dir contains config.json
// model/config.json
{ "_class_name": "UNet2DConditionModel", "in_channels": 4 }

cfg_path = model_path / "config.json"
if not cfg_path.is_file():
    raise FileNotFoundError(f"Model directory missing config.json: {model_path}")

def config_json_present(model_path) -> bool:
    import json
    p = model_path / "config.json"
    return p.is_file() and bool(json.loads(p.read_text()))

try:
    cls = loader.get_hf_load_class(model_path)
except ValueError as e:
    if "config.json" in str(e):
        re_download_model(model_path)  # restore missing metadata files
    else:
        raise

When it happens

Trigger: get_hf_load_class hits a KeyError while resolving class names from config.json — e.g. the config file itself is absent (lookup raised KeyError) or `_class_name`/`architectures` values point to entries missing from the loader's expected structure.

Common situations: Interrupted downloads leaving models without config.json; models converted with tools that drop metadata files; users pointing the loader at a checkpoint root that only contains weights.

Related errors

• missing pytorch_lora_weights.bin or pytorch_lora_weights.saf
• Unable to decipher Load Class based on given config.json
• No weight files found for this model
• str(e)
• Multiuser mode is disabled. Authentication is not required i

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/59427aa046ca2b64. Report an issue: GitHub.