huggingface/transformers · error · ValueError

WeightConverter requires at least one operation.

Error message

WeightConverter requires at least one operation.

What it means

Raised by WeightConverter.__init__ (core_model_loading.py:1154). A WeightConverter is defined by its list of tensor operations (the actual conversion math); an empty operations list means there is nothing to convert, which almost always indicates the caller passed an uninitialized/empty list or forgot the argument. The constructor validates this after the cardinality check and refuses to build an operation-less converter.

Source

Thrown at src/transformers/core_model_loading.py:1154

class WeightConverter(WeightTransform):
    __slots__ = ("operations",)

    def __init__(
        self, source_patterns: str | list[str], target_patterns: str | list[str], operations: list[ConversionOps]
    ):
        super().__init__(source_patterns, target_patterns)
        self.operations: list[ConversionOps] = operations

        if bool(len(self.source_patterns) - 1) + bool(len(self.target_patterns) - 1) >= 2:
            # We allow many-to-many only if we use an internal operation that can handle it
            if not any(isinstance(op, _INTERNAL_MANY_TO_MANY_CONVERSIONS) for op in self.operations):
                raise ValueError(
                    f"source keys={self.source_patterns}, target_patterns={self.target_patterns} but you can only have one to many, one to one or many to one."
                )
        if not self.operations:
            raise ValueError("WeightConverter requires at least one operation.")

    def convert(
        self,
        layer_name: str,
        model=None,
        config=None,
        hf_quantizer=None,
        loading_info: LoadStateDictInfo | None = None,
    ):
        # Collect the tensors here - we use a new dictionary to avoid keeping them in memory in the internal
        # attribute during the whole process
        collected_tensors = self.materialize_tensors()

        for op in self.operations:
            with log_conversion_errors(layer_name, loading_info, (len(collected_tensors), layer_name), op):
                collected_tensors = op.convert(
                    collected_tensors,
                    source_patterns=self.source_patterns,

View on GitHub (pinned to a597f97485)

Solutions

1. If you only need key renaming with no tensor math, use WeightTransform/GroupWeightRename/PrefixChange instead of WeightConverter.
2. If tensor conversion is intended, pass at least one op (e.g. Identity-like op or the real conversion) in operations.
3. In programmatic builders, assert the ops list is non-empty before constructing to surface the upstream logic error.

Example fix

# before
WeightConverter(source_patterns=[r'blk.*'], target_patterns=[r'layers.*'], operations=ops)  # ops == []

# after: pure rename -> WeightTransform
WeightTransform(source_patterns=[r'blk.*'], target_patterns=[r'layers.*'])

assert isinstance(operations, list) and len(operations) > 0, (
    'WeightConverter needs >=1 operation; use WeightTransform for pure renames'
)

def has_operations(operations) -> bool:
    return bool(operations)

When it happens

Trigger: WeightConverter(source_patterns=[...], target_patterns=[...], operations=[]) — e.g. operations were built conditionally and the condition never fired, leaving an empty list; or a refactor left the default empty list in place.

Common situations: Programmatic recipe builders that accumulate ops in a loop which never executes (empty config section, wrong filter), or copy-paste where the operations argument was dropped. If you only need a rename, use WeightTransform (no operations) instead of WeightConverter.

Related errors

• GroupWeightRename requires N:N length matching, but found le
• You must provide only one of `prefix_to_add` and `prefix_to_
• source keys={self.source_patterns}, target_patterns={self.ta
• Multiple different capturing groups found in target_patterns
• Source pattern '{pattern}' contains \\1 backreference, but n

AI-assisted analysis of huggingface/transformers@a597f97485 (2026-08-14). Data as JSON: /api/errors/fd52d12afb660a6d. Report an issue: GitHub.