tensorflow/models · error · ValueError

Unknown `attention_type` {}.

Error message

Unknown `attention_type` {}.

What it means

Error "Unknown `attention_type` {}." thrown in tensorflow/models.

Source

Thrown at official/nlp/modeling/networks/xlnet_base.py:318

      means no clamping.
    bi_data: bool, whether to use bidirectional input pipeline. Usually set to
      True during pretraining and False during finetuning.
    dtype: the dtype of the encoding.

  Returns:
    A Tensor, representing the position encoding.

  """
  freq_seq = tf.range(0, hidden_size, 2.0)
  if dtype is not None and dtype != tf.float32:
    freq_seq = tf.cast(freq_seq, dtype=dtype)

  if attention_type == "bi":
    beg, end = total_length, -seq_length
  elif attention_type == "uni":
    beg, end = total_length, -1
  else:
    raise ValueError("Unknown `attention_type` {}.".format(attention_type))

  if bi_data:
    forward_position_sequence = tf.range(beg, end, -1.0)
    backward_position_sequence = tf.range(-beg, -end, 1.0)

    if dtype is not None and dtype != tf.float32:
      forward_position_sequence = tf.cast(forward_position_sequence,
                                          dtype=dtype)
      backward_position_sequence = tf.cast(backward_position_sequence,
                                           dtype=dtype)

    if clamp_length > 0:
      forward_position_sequence = tf.clip_by_value(
          forward_position_sequence,
          -clamp_length,
          clamp_length)
      backward_position_sequence = tf.clip_by_value(
          backward_position_sequence,

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/modeling/networks/xlnet_base.py:318 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/85cea32231303370. Report an issue: GitHub.