tensorflow/models · error · ValueError

Attribute type {head.type} not supported.

Error message

Attribute type {head.type} not supported.

What it means

Error "Attribute type {head.type} not supported." thrown in tensorflow/models.

Source

Thrown at official/vision/tasks/retinanet.py:210

        y_true_att = loss_utils.multi_level_flatten(
            labels['attribute_targets'][head.name], last_dim=None
        )
        y_true_att = tf.one_hot(y_true_att, head.size)
        y_pred_att = loss_utils.multi_level_flatten(
            outputs['attribute_outputs'][head.name], last_dim=head.size
        )
        cls_loss_fn = focal_loss.FocalLoss(
            alpha=params.losses.focal_loss_alpha,
            gamma=params.losses.focal_loss_gamma,
            reduction=tf_keras.losses.Reduction.SUM,
        )
        att_loss = cls_loss_fn(
            y_true=y_true_att,
            y_pred=y_pred_att,
            sample_weight=box_sample_weight,
        )
      else:
        raise ValueError(f'Attribute type {head.type} not supported.')
      attribute_loss += att_loss

    return attribute_loss

  def build_losses(
      self,
      outputs: Mapping[str, Any],
      labels: Mapping[str, Any],
      aux_losses: Optional[Any] = None,
  ):
    """Build RetinaNet losses."""
    params = self.task_config
    attribute_heads = self.task_config.model.head.attribute_heads

    cls_loss_fn = focal_loss.FocalLoss(
        alpha=params.losses.focal_loss_alpha,
        gamma=params.losses.focal_loss_gamma,
        reduction=tf_keras.losses.Reduction.SUM)

View on GitHub (pinned to e006f5f0d5)

Solutions

1. Use a supported attribute head type (e.g. 'classification' or 'regression') in the RetinaNet task config.
2. Check the attribute head 'type' field for typos.

When it happens

Trigger: Thrown at official/vision/tasks/retinanet.py:210 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/63f082bce77689a2. Report an issue: GitHub.