huggingface/transformers · error · ValueError

The `layer_types` entries must be in {ALLOWED_LAYER_TYPES}

Error message

The `layer_types` entries must be in {ALLOWED_LAYER_TYPES}

What it means

ValueError from the deprecated free function layer_type_validation (warned for removal in v5.20; use PreTrainedConfig.validate_layer_type instead): a value in layer_types is not a member of ALLOWED_LAYER_TYPES (e.g. 'dense'/'linear' attention variants).

Source

Thrown at src/transformers/configuration_utils.py:1489

PreTrainedConfig.push_to_hub = copy_func(PreTrainedConfig.push_to_hub)
if PreTrainedConfig.push_to_hub.__doc__ is not None:
    PreTrainedConfig.push_to_hub.__doc__ = PreTrainedConfig.push_to_hub.__doc__.format(
        object="config", object_class="AutoConfig", object_files="configuration file"
    )


# The alias is only here for BC - we did not have the correct CamelCasing before
PretrainedConfig = PreTrainedConfig


def layer_type_validation(layer_types: list[str], num_hidden_layers: int | None = None, attention: bool = True):
    logger.warning(
        "`layer_type_validation` is deprecated and will be removed in v5.20. "
        "Use `PreTrainedConfig.validate_layer_type` instead"
    )

    if not all(layer_type in ALLOWED_LAYER_TYPES for layer_type in layer_types):
        raise ValueError(f"The `layer_types` entries must be in {ALLOWED_LAYER_TYPES}")
    if num_hidden_layers is not None and num_hidden_layers != len(layer_types):
        raise ValueError(
            f"`num_hidden_layers` ({num_hidden_layers}) must be equal to the number of layer types "
            f"({len(layer_types)})"
        )

View on GitHub (pinned to a597f97485)

Solutions

1. Use only strings from transformers.configuration_utils.ALLOWED_LAYER_TYPES (check the constant in your version).
2. Upgrade Transformers if the layer type is valid in a newer release.
3. Migrate to the instance API: config.validate_layer_type(layer_types, num_hidden_layers).

Example fix

// before
layer_type_validation(["dense", "full"], num_hidden_layers=2)  # ValueError

// after
layer_type_validation(["dense", "linear"], num_hidden_layers=2)

from transformers.configuration_utils import ALLOWED_LAYER_TYPES
bad = [t for t in layer_types if t not in ALLOWED_LAYER_TYPES]
if bad:
    raise ValueError(f"unsupported layer types {bad}; allowed: {ALLOWED_LAYER_TYPES}")

def layer_types_are_valid(layer_types: list[str]) -> bool:
    from transformers.configuration_utils import ALLOWED_LAYER_TYPES
    return all(t in ALLOWED_LAYER_TYPES for t in layer_types)

When it happens

Trigger: Calling layer_type_validation(["dense", "sliding", "full"], ...) where 'sliding'/'full' are not in ALLOWED_LAYER_TYPES, or configs carrying a custom layer type string not yet supported by the installed version.

Common situations: New architectures adding a layer type before the installed Transformers version knows it (version skew); typos like 'Dense' with wrong casing.

Related errors

• `num_hidden_layers` ({num_hidden_layers}) must be equal to t
• The `{layer_types}` entries must be in {allowed_types} but g
• `num_hidden_layers` ({self.num_hidden_layers}) must be equal
• Layer type '{layer_idx}' not found in config.layer_types: {l
• out_indices must be a list, got {type(self._out_indices)}

AI-assisted analysis of huggingface/transformers@a597f97485 (2026-08-14). Data as JSON: /api/errors/aa5df2a719904b55. Report an issue: GitHub.