tensorflow/models · error · ValueError

max_num_eval_detections must be an integer.

Error message

max_num_eval_detections must be an integer.

What it means

Error "max_num_eval_detections must be an integer." thrown in tensorflow/models.

Source

Thrown at official/vision/evaluation/coco_evaluator.py:98

    if annotation_file:
      if annotation_file.startswith('gs://'):
        _, local_val_json = tempfile.mkstemp(suffix='.json')
        tf.io.gfile.remove(local_val_json)

        tf.io.gfile.copy(annotation_file, local_val_json)
        atexit.register(tf.io.gfile.remove, local_val_json)
      else:
        local_val_json = annotation_file
      self._coco_gt = coco_utils.COCOWrapper(
          eval_type=('mask' if include_mask else 'box'),
          annotation_file=local_val_json)
    self._annotation_file = annotation_file
    self._include_mask = include_mask
    self._include_keypoint = include_keypoint
    self._per_category_metrics = per_category_metrics
    if max_num_eval_detections is None or not isinstance(
        max_num_eval_detections, int):
      raise ValueError('max_num_eval_detections must be an integer.')
    self._metric_names = [
        'AP', 'AP50', 'AP75', 'APs', 'APm', 'APl', 'ARmax1', 'ARmax10',
        f'ARmax{max_num_eval_detections}', 'ARs', 'ARm', 'ARl'
    ]
    self.max_num_eval_detections = max_num_eval_detections
    self._required_prediction_fields = [
        'source_id', 'num_detections', 'detection_classes', 'detection_scores',
        'detection_boxes'
    ]
    self._need_rescale_bboxes = need_rescale_bboxes
    self._need_rescale_keypoints = need_rescale_keypoints
    if self._need_rescale_bboxes or self._need_rescale_keypoints:
      self._required_prediction_fields.append('image_info')
    self._required_groundtruth_fields = [
        'source_id', 'height', 'width', 'classes', 'boxes'
    ]
    if self._include_mask:
      mask_metric_names = ['mask_' + x for x in self._metric_names]

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/vision/evaluation/coco_evaluator.py:98 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/0609a92a3232328d. Report an issue: GitHub.