tensorflow/models · error · ValueError

Unknown crop variant: {variant!r}

Error message

Unknown crop variant: {variant!r}

What it means

Error "Unknown crop variant: {variant!r}" thrown in tensorflow/models.

Source

Thrown at official/projects/waste_identification_ml/data_generation/auto_labeler_pipeline/segmentation.py:285

  Returns:
      A PIL image for the requested variant, or ``None`` for degenerate
      boxes in the ``'raw'`` variant.

  Raises:
      ValueError: If ``variant`` is not one of the allowed values.
  """
  if variant == _RAW_VARIANT:
    return sam3_inference_utils.crop_raw_masked_image(image_array, mask, box)
  if variant == _BLACK_BACKGROUND_VARIANT:
    return sam3_inference_utils.crop_masked_image(
        image_array, mask, box, size=crop_size
    )
  if variant == _IMAGENET_MEAN_BACKGROUND_VARIANT:
    return sam3_inference_utils.crop_with_mean_background_blend(
        image_array, mask, box, size=crop_size
    )
  raise ValueError(f"Unknown crop variant: {variant!r}")


def generate_selected_crops(
    image: Image.Image,
    state: dict[str, Any],
    score_threshold: float,
    crop_size: tuple[int, int],
    variants: tuple[str, ...],
) -> list[tuple[int, dict[str, Any]]]:
  """Generates only the crop variants requested in ``variants``.

  Skips the work of building unused crop variants. Mask hole-filling is
  performed once per detection and reused across variants.

  Args:
      image: Input RGB PIL image.
      state: SAM output dict with ``'masks'``, ``'boxes'``, ``'scores'`` keys.
      score_threshold: Minimum confidence score to include a detection.

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/projects/waste_identification_ml/data_generation/auto_labeler_pipeline/segmentation.py:285 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/58bca88c9b575b85. Report an issue: GitHub.