mudler/LocalAI · error · ValueError

Could not find base class '{base_class_name}' in diffusers

Error message

Could not find base class '{base_class_name}' in diffusers

What it means

Raised by the diffusers dynamic loader when building the pipeline registry: the requested base class name cannot be found on the top-level diffusers module nor in the 'schedulers', 'models', 'pipelines' submodules. The loader then cannot enumerate subclasses to populate the registry.

Source

Thrown at backend/python/diffusers/diffusers_dynamic_loader.py:180

    import diffusers

    # Try to get the base class from diffusers
    base_class = None
    try:
        base_class = getattr(diffusers, base_class_name)
    except AttributeError:
        # Try to find in submodules
        for submodule in ['schedulers', 'models', 'pipelines']:
            try:
                module = importlib.import_module(f'diffusers.{submodule}')
                if hasattr(module, base_class_name):
                    base_class = getattr(module, base_class_name)
                    break
            except (ImportError, ModuleNotFoundError):
                continue

    if base_class is None:
        raise ValueError(f"Could not find base class '{base_class_name}' in diffusers")

    registry: Dict[str, Type] = {}

    # Include base class if requested
    if include_base:
        registry[base_class_name] = base_class

    # Scan diffusers module for subclasses
    for attr_name in dir(diffusers):
        try:
            attr = getattr(diffusers, attr_name)
            if (isinstance(attr, type) and
                issubclass(attr, base_class) and
                (include_base or attr is not base_class)):
                registry[attr_name] = attr
        except (ImportError, AttributeError, TypeError, RuntimeError, ModuleNotFoundError):
            continue

View on GitHub (pinned to 44413a9d06)

Solutions

1. Print [c for c in dir(diffusers) if 'Pipeline' in c] against the installed diffusers to find the current class name.
2. Upgrade/downgrade diffusers to the version the backend code was written against (check the backend image's requirements pin).
3. Use the canonical base class 'DiffusionPipeline' unless a specific subclass registry is required.
4. Fix the typo/case of base_class_name at the call site.

import diffusers
assert hasattr(diffusers, base_class_name) or any(
    hasattr(importlib.import_module(f'diffusers.{m}'), base_class_name)
    for m in ('schedulers', 'models', 'pipelines')
), f"base class {base_class_name} absent from diffusers {diffusers.__version__}"

def base_class_exists(name: str) -> bool:
    import diffusers
    if hasattr(diffusers, name):
        return True
    return any(
        hasattr(importlib.import_module(f'diffusers.{m}'), name)
        for m in ('schedulers', 'models', 'pipelines')
    )

try:
    registry = _discover_pipelines(base_class_name)
except ValueError as e:
    logger.error("%s (diffusers %s)", e, diffusers.__version__)
    registry = _discover_pipelines("DiffusionPipeline")  # only if fallback acceptable

When it happens

Trigger: get_pipeline_registry (or a caller) requests a base class name that does not exist in the installed diffusers version — e.g. a renamed or removed class, or a typo in the base_class_name argument.

Common situations: diffusers version drift (class renamed upstream, e.g. scheduler refactorings), a backend image pinned to an older diffusers while the code expects a newer class, or custom pipeline loading paths passing an arbitrary string as base class.

Related errors

• Unknown pipeline: {class_name}
• No job ID returned from server
• Invalid scheduler '{'k_' if is_karras else ''}{name}'
• Failed to load pipeline '{effective_pipeline_type}': {e}\nAv
• Unknown pipeline class '{class_name}'. Available pipelines:

AI-assisted analysis of mudler/LocalAI@44413a9d06 (2026-08-15). Data as JSON: /api/errors/0d1f41e9bc111dee. Report an issue: GitHub.