tensorflow/models · error · ValueError

PointPillars model needs attribute heads.

Error message

PointPillars model needs attribute heads.

What it means

Error "PointPillars model needs attribute heads." thrown in tensorflow/models.

Source

Thrown at official/projects/pointpillars/modeling/models.py:72

    self._backbone = backbone
    self._decoder = decoder
    self._head = head
    self._detection_generator = detection_generator
    self._min_level = min_level
    self._max_level = max_level
    self._image_size = image_size
    self._anchor_sizes = anchor_sizes

  def generate_outputs(
      self,
      raw_scores: Dict[str, tf.Tensor],
      raw_boxes: Dict[str, tf.Tensor],
      raw_attributes: Dict[str, Dict[str, tf.Tensor]],
      image_shape: Optional[tf.Tensor] = None,
      anchor_boxes: Optional[Mapping[str, tf.Tensor]] = None,
      generate_detections: bool = False) -> Mapping[str, Any]:
    if not raw_attributes:
      raise ValueError('PointPillars model needs attribute heads.')
    # Clap heading to [-pi, pi]
    if 'heading' in raw_attributes:
      raw_attributes['heading'] = utils.clip_heading(raw_attributes['heading'])

    outputs = {
        'cls_outputs': raw_scores,
        'box_outputs': raw_boxes,
        'attribute_outputs': raw_attributes,
    }
    # Cast raw prediction to float32 for loss calculation.
    outputs = tf.nest.map_structure(lambda x: tf.cast(x, tf.float32), outputs)
    if not generate_detections:
      return outputs

    if image_shape is None:
      raise ValueError('Image_shape should not be None for evaluation.')
    if anchor_boxes is None:
      # Generate anchors if needed.

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/projects/pointpillars/modeling/models.py:72 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/ec13f8cca3cc5133. Report an issue: GitHub.