tensorflow/models · error · ValueError

reuse_attention_scores cannot be None when reuse_attention !

Error message

reuse_attention_scores cannot be None when reuse_attention != 0.

What it means

Error "reuse_attention_scores cannot be None when reuse_attention != 0." thrown in tensorflow/models.

Source

Thrown at official/nlp/modeling/layers/reuse_transformer.py:307

      Attention scores if return_attention_scores is true.
    """
    if isinstance(inputs, (list, tuple)):
      if len(inputs) == 2:
        input_tensor, attention_mask = inputs
        reuse_attention_scores = None
      elif len(inputs) == 3:
        input_tensor, attention_mask, reuse_attention_scores = inputs
      else:
        raise ValueError("Unexpected inputs to %s with length at %d" %
                         (self.__class__, len(inputs)))
    else:
      input_tensor, attention_mask, reuse_attention_scores = (inputs, None,
                                                              None)

    key_value = None

    if self._reuse_attention != 0 and reuse_attention_scores is None:
      raise ValueError(
          "reuse_attention_scores cannot be None when reuse_attention != 0.")

    if self._output_range:
      if self._norm_first:
        source_tensor = input_tensor[:, 0:self._output_range, :]
        input_tensor = self._attention_layer_norm(input_tensor)
        if key_value is not None:
          key_value = self._attention_layer_norm(key_value)
      target_tensor = input_tensor[:, 0:self._output_range, :]
      if attention_mask is not None:
        attention_mask = attention_mask[:, 0:self._output_range, :]
      if reuse_attention_scores is not None:
        reuse_attention_scores = reuse_attention_scores[:, :,
                                                        0:self._output_range, :]
    else:
      if self._norm_first:
        source_tensor = input_tensor
        input_tensor = self._attention_layer_norm(input_tensor)

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/nlp/modeling/layers/reuse_transformer.py:307 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/eaff775d4aaecd46. Report an issue: GitHub.