tensorflow/models · error · ValueError

Unrecognized `input_type`

Error message

Unrecognized `input_type`

What it means

Error "Unrecognized `input_type`" thrown in tensorflow/models.

Source

Thrown at official/vision/serving/export_utils.py:53

    input_name: A `str` to set the input image name in the signature, if None,
      a default name `inputs` will be used.
  Returns:
    tf.TensorSpec of the input tensor.
  """
  if input_type == 'image_tensor':
    input_signature = tf.TensorSpec(
        shape=[batch_size] + [None] * len(input_image_size) + [num_channels],
        dtype=tf.uint8, name=input_name)
  elif input_type in ['image_bytes', 'serve_examples', 'tf_example']:
    input_signature = tf.TensorSpec(
        shape=[batch_size], dtype=tf.string, name=input_name)
  elif input_type == 'tflite':
    input_signature = tf.TensorSpec(
        shape=[1] + input_image_size + [num_channels],
        dtype=tf.float32,
        name=input_name)
  else:
    raise ValueError('Unrecognized `input_type`')
  return input_signature


def decode_image(encoded_image_bytes: str,
                 input_image_size: List[int],
                 num_channels: int = 3,) -> tf.Tensor:
  """Decodes an image bytes to an image tensor.

  Use `tf.image.decode_image` to decode an image if input is expected to be 2D
  image; otherwise use `tf.io.decode_raw` to convert the raw bytes to tensor
  and reshape it to desire shape.

  Args:
    encoded_image_bytes: An encoded image string to be decoded.
    input_image_size: List[int] for the desired input size. This will be used to
      infer whether the image is 2d or 3d.
    num_channels: `int` for number of image channels.

View on GitHub (pinned to e006f5f0d5)

Solutions

1. Set input_type to 'image_tensor', 'image_bytes', or 'tf_example'.
2. Check the serving input_type config for typos.

When it happens

Trigger: Thrown at official/vision/serving/export_utils.py:53 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/5f5acfc86484da87. Report an issue: GitHub.