tensorflow/models · error · ValueError

Not supported loss mode: {loss_mode}

Error message

Not supported loss mode: {loss_mode}

What it means

Error "Not supported loss mode: {loss_mode}" thrown in tensorflow/models.

Source

Thrown at official/projects/unified_detector/modeling/universal_detector.py:855

        tf.reduce_sum(pointwise_loss * pos_mask) /
        (tf.reduce_sum(pos_mask) + EPSILON))
    # neg
    neg_mask = gt_affinity_mask * (1. - gt_affinity[:, 0])
    neg_loss = (
        tf.reduce_sum(pointwise_loss * neg_mask) /
        (tf.reduce_sum(neg_mask) + EPSILON))
    loss = 0.25 * pos_loss + 0.75 * neg_loss
  elif loss_mode == "focal":
    alpha_wt = fl_alpha * gt_affinity + (1. - fl_alpha) * (1. - gt_affinity)
    prob_pos = tf.math.sigmoid(affinity / tau)
    pt = prob_pos * gt_affinity + (1. - prob_pos) * (1. - gt_affinity)
    fl_loss_pw = tf.stop_gradient(
        alpha_wt * tf.pow(1. - pt, fl_gamma))[:, 0] * pointwise_loss
    loss = (
        tf.reduce_sum(fl_loss_pw * gt_affinity_mask) /
        (tf.reduce_sum(gt_affinity_mask) + EPSILON))
  else:
    raise ValueError(f"Not supported loss mode: {loss_mode}")

  loss_dict["loss_para"] = loss


def _mask_id_xent_loss(loss_dict: Dict[str, Any], labels: Dict[str, Any],
                       outputs: Dict[str, Any]):
  """Mask ID loss.

  This method adds the mask ID loss term to loss_dict directly.

  Args:
    loss_dict: A dictionary for the loss. The values are loss scalars.
    labels: The label dictionary.
    outputs: The output dictionary.
  """
  # (B, N, H, W)
  mask_gt = labels["masks"]
  # B, H, W, N

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/projects/unified_detector/modeling/universal_detector.py:855 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/f6397b146ce1d757. Report an issue: GitHub.