tensorflow/models · error · ValueError

encoder_network's output should be either a list or a dict,

Error message

encoder_network's output should be either a list or a dict, but got %s

What it means

Error "encoder_network's output should be either a list or a dict, but got %s" thrown in tensorflow/models.

Source

Thrown at official/projects/edgetpu/nlp/modeling/pretrainer.py:76

      classification_heads: Optional[List[tf_keras.layers.Layer]] = None,
      customized_masked_lm: Optional[tf_keras.layers.Layer] = None,
      name: str = 'bert',
      **kwargs):

    inputs = copy.copy(encoder_network.inputs)
    outputs = {}
    encoder_network_outputs = encoder_network(inputs)
    if isinstance(encoder_network_outputs, list):
      outputs['pooled_output'] = encoder_network_outputs[1]
      if isinstance(encoder_network_outputs[0], list):
        outputs['encoder_outputs'] = encoder_network_outputs[0]
        outputs['sequence_output'] = encoder_network_outputs[0][-1]
      else:
        outputs['sequence_output'] = encoder_network_outputs[0]
    elif isinstance(encoder_network_outputs, dict):
      outputs = encoder_network_outputs
    else:
      raise ValueError('encoder_network\'s output should be either a list '
                       'or a dict, but got %s' % encoder_network_outputs)

    masked_lm_positions = tf_keras.layers.Input(
        shape=(None,), name='masked_lm_positions', dtype=tf.int32)
    inputs.append(masked_lm_positions)
    masked_lm_layer = customized_masked_lm or layers.MaskedLM(
        embedding_table=encoder_network.get_embedding_table(),
        activation=mlm_activation,
        initializer=mlm_initializer,
        name='cls/predictions')
    sequence_output = outputs['sequence_output']
    outputs['mlm_logits'] = masked_lm_layer(
        sequence_output, masked_positions=masked_lm_positions)

    classification_head_layers = classification_heads or []
    for cls_head in classification_head_layers:
      cls_outputs = cls_head(sequence_output)
      if isinstance(cls_outputs, dict):

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/projects/edgetpu/nlp/modeling/pretrainer.py:76 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/54825ab966a0dc87. Report an issue: GitHub.