tensorflow/models · error · ValueError

Exactly one of `bert_config` and `encoder_config` can be spe

Error message

Exactly one of `bert_config` and `encoder_config` can be specified, but got %s and %s

What it means

Error "Exactly one of `bert_config` and `encoder_config` can be specified, but got %s and %s" thrown in tensorflow/models.

Source

Thrown at official/nlp/tools/export_tfhub_lib.py:101

  """Creates the model to export and the model to restore the checkpoint.

  Args:
    bert_config: A legacy `BertConfig` to create a `BertEncoder` object. Exactly
      one of encoder_config and bert_config must be set.
    encoder_config: An `EncoderConfig` to create an encoder of the configured
      type (`BertEncoder` or other).
    with_mlm: A bool to control the second component of the result. If True,
      will create a `BertPretrainerV2` object; otherwise, will create a
      `BertEncoder` object.

  Returns:
    A Tuple of (1) a Keras model that will be exported, (2) a `BertPretrainerV2`
    object or `BertEncoder` object depending on the value of `with_mlm`
    argument, which contains the first model and will be used for restoring
    weights from the checkpoint.
  """
  if (bert_config is not None) == (encoder_config is not None):
    raise ValueError("Exactly one of `bert_config` and `encoder_config` "
                     "can be specified, but got %s and %s" %
                     (bert_config, encoder_config))

  if bert_config is not None:
    encoder = get_bert_encoder(bert_config)
  else:
    encoder = encoders.build_encoder(encoder_config)

  # Convert from list of named inputs to dict of inputs keyed by name.
  # Only the latter accepts a dict of inputs after restoring from SavedModel.
  if isinstance(encoder.inputs, list) or isinstance(encoder.inputs, tuple):
    encoder_inputs_dict = {x.name: x for x in encoder.inputs}
  else:
    # encoder.inputs by default is dict for BertEncoderV2.
    encoder_inputs_dict = encoder.inputs
  encoder_output_dict = encoder(encoder_inputs_dict)
  # For interchangeability with other text representations,
  # add "default" as an alias for BERT's whole-input reptesentations.

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/tools/export_tfhub_lib.py:101 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of tensorflow/models@e006f5f0d5 (2026-08-24). Data as JSON: /api/errors/89960e95f54aa2ed. Report an issue: GitHub.