tensorflow/models · error · ValueError

Attribute learning is only supported for NMSv1 but NMS {} is

Error message

Attribute learning is only supported for NMSv1 but NMS {} is used.

What it means

Error "Attribute learning is only supported for NMSv1 but NMS {} is used." thrown in tensorflow/models.

Source

Thrown at official/vision/modeling/layers/detection_generator.py:1515

    decoded_results = {
        'decoded_boxes': boxes,
        'decoded_box_scores': scores,
        'decoded_box_attributes': attributes,
    }

    if not self._config_dict['apply_nms']:
      return decoded_results

    # Optionally force the NMS to run on CPU.
    if self._config_dict['use_cpu_nms']:
      nms_context = tf.device('cpu:0')
    else:
      nms_context = contextlib.nullcontext()

    with nms_context:
      if raw_attributes and (self._config_dict['nms_version'] != 'v1'):
        raise ValueError(
            'Attribute learning is only supported for NMSv1 but NMS {} is used.'
            .format(self._config_dict['nms_version'])
        )
      if self._config_dict['nms_version'] == 'batched':
        (nmsed_boxes, nmsed_scores, nmsed_classes, valid_detections) = (
            _generate_detections_batched(
                boxes,
                scores,
                self._config_dict['pre_nms_score_threshold'],
                self._config_dict['nms_iou_threshold'],
                self._config_dict['max_num_detections'],
            )
        )
        # Set `nmsed_attributes` to None for batched NMS.
        nmsed_attributes = {}
      elif self._config_dict['nms_version'] == 'v1':
        (
            nmsed_boxes,

View on GitHub (pinned to e006f5f0d5)

Solutions

  1. Set nms_version to 1 when attribute learning is enabled.
  2. Disable attribute learning heads if you need a different NMS version.

When it happens

Trigger: Thrown at official/vision/modeling/layers/detection_generator.py:1515 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/0fad7f4e214e774b. Report an issue: GitHub.