tensorflow/models · error · ValueError

'num_positives' was provided but it was a negative number.

Error message

'num_positives' was provided but it was a negative number.

What it means

Argument check in AveragePrecisionCalculator.accumulate: the optional num_positives override was supplied but is negative. num_positives lets callers account for true positives absent from the passed arrays; a negative count is nonsensical and would inflate precision, so it is rejected before any computation.

Solutions

1. Pass num_positives >= 0, or omit it (None) to infer from actuals
2. Check the upstream counting logic for a sign error or subtraction bug
3. Coerce numpy scalars to int if a non-int numeric sneaks in from metrics code

When it happens

Trigger: Thrown at official/projects/yt8m/eval_utils/average_precision_calculator.py:118 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/3767f8918b3cdfea. Report an issue: GitHub.

Appendix: source

Thrown at official/projects/yt8m/eval_utils/average_precision_calculator.py:118

    Args:
      predictions: a list storing the prediction scores.
      actuals: a list storing the ground truth labels. Any value larger than 0
        will be treated as positives, otherwise as negatives. num_positives = If
        the 'predictions' and 'actuals' inputs aren't complete, then it's
        possible some true positives were missed in them. In that case, you can
        provide 'num_positives' in order to accurately track recall.
      num_positives: number of positive examples.

    Raises:
      ValueError: An error occurred when the format of the input is not the
      numpy 1-D array or the shape of predictions and actuals does not match.
    """
    if len(predictions) != len(actuals):
      raise ValueError("the shape of predictions and actuals does not match.")

    if num_positives is not None:
      if not isinstance(num_positives, numbers.Number) or num_positives < 0:  # pyrefly: ignore[unsupported-operation]
        raise ValueError(
            "'num_positives' was provided but it was a negative number.")

    if num_positives is not None:
      self._total_positives += num_positives
    else:
      self._total_positives += numpy.size(
          numpy.where(numpy.array(actuals) > 1e-5))
    topk = self._top_n
    heap = self._heap

    for i in range(numpy.size(predictions)):
      if topk is None or len(heap) < topk:
        heapq.heappush(heap, (predictions[i], actuals[i]))
      else:
        if predictions[i] > heap[0][0]:  # heap[0] is the smallest
          heapq.heappop(heap)
          heapq.heappush(heap, (predictions[i], actuals[i]))

View on GitHub (pinned to e006f5f0d5)