invoke-ai/InvokeAI · error · ValueError

missing _class_name or architectures field

Error message

missing _class_name or architectures field

What it means

NotAMatchError raised by get_class_name_from_config_dict_or_raise when a successfully loaded config dict contains neither a '_class_name' key (diffusers-style configs) nor an 'architectures' key (transformers-style configs), so no architecture marker can be extracted. This is wrapped into NotAMatchError with the 'unable to determine class name' message (1039), whose cause chain shows this ValueError.

Source

Thrown at invokeai/backend/model_manager/configs/identification_utils.py:104

    Returns:
        The class name from the config file.

    Raises:
        NotAMatch if the config file is missing or does not contain a valid class name.
    """

    if not isinstance(config, dict):
        config = get_config_dict_or_raise(config)

    try:
        if "_class_name" in config:
            # This is a diffusers-style config
            config_class_name = config["_class_name"]
        elif "architectures" in config:
            # This is a transformers-style config
            config_class_name = config["architectures"][0]
        else:
            raise ValueError("missing _class_name or architectures field")
    except Exception as e:
        raise NotAMatchError(f"unable to determine class name from config file: {config}") from e

    if not isinstance(config_class_name, str):
        raise NotAMatchError(f"_class_name or architectures field is not a string: {config_class_name}")

    return config_class_name


def raise_for_class_name(config: Path | set[Path] | dict[str, Any], class_name: str | set[str]) -> None:
    """Get the class name from the config file and raise NotAMatch if it is not in the expected set.

    Args:
        config_path: The path to the config file, or a set of paths to try.
        class_name: The expected class name, or a set of expected class names.

    Raises:
        NotAMatch if the class name is not in the expected set.

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Add "architectures": ["Gemma2ForCausalLM"] (or the appropriate class) to config.json, or "_class_name" for diffusers-style configs
2. Re-download config.json from the original HuggingFace repo instead of a hand-made one
3. Point the importer at the correct config file — you may be reading a secondary config (e.g. tokenizer_config.json-style file) that lacks these keys

Example fix

// before
{ "hidden_size": 2304, "model_type": "gemma2" }
// after
{ "architectures": ["Gemma2ForCausalLM"], "hidden_size": 2304, "model_type": "gemma2" }

import json
from pathlib import Path

def config_has_class_name(model_dir: str | Path) -> bool:
    p = Path(model_dir) / "config.json"
    try:
        cfg = json.loads(p.read_text(encoding="utf-8"))
    except (OSError, json.JSONDecodeError):
        return False
    archs = cfg.get("architectures")
    return isinstance(cfg.get("_class_name"), str) or (
        isinstance(archs, list) and len(archs) > 0 and isinstance(archs[0], str)
    )

def extract_class_name(cfg: dict) -> str | None:
    if isinstance(cfg.get("_class_name"), str):
        return cfg["_class_name"]
    archs = cfg.get("architectures")
    if isinstance(archs, list) and archs and isinstance(archs[0], str):
        return archs[0]
    return None

try:
    import_model(model_dir)
except NotAMatchError as e:
    if "unable to determine class name" in str(e):
        print("config.json lacks _class_name/architectures — restore the original from the HF repo")

When it happens

Trigger: get_class_name_from_config_dict_or_raise / raise_for_class_name / from_model_on_disk receiving a config dict (e.g. model_index.json, custom config.json, or a hand-written JSON) lacking both keys — commonly a bare {"model_type": ...}-only transformers config or an empty {} dict.

Common situations: Custom model exports that omit _class_name; older or minimal transformers configs without 'architectures'; user-authored placeholder config.json; configs trimmed by download managers that only keep model_type.

Related errors

• unable to load config file(s): {problems}
• unable to determine class name from config file: {config}
• Gemini response payload was not a JSON object
• Gemma2 hidden_size {hidden_size} is incompatible with PiD, w
• directory does not contain Gemma2 tokenizer files (tokenizer

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/801c66576eb0d2bd. Report an issue: GitHub.