tensorflow/models · error · ValueError

Task {} not supported.

Error message

Task {} not supported.

What it means

Error "Task {} not supported." thrown in tensorflow/models.

Source

Thrown at official/vision/serving/export_tflite_lib.py:57

    A tf.data.Dataset instance.

  Raises:
    ValueError: If task is not supported.
  """
  if task is None:
    if isinstance(params.task,
                  configs.image_classification.ImageClassificationTask):

      task = tasks.image_classification.ImageClassificationTask(params.task)
    elif isinstance(params.task, configs.retinanet.RetinaNetTask):
      task = tasks.retinanet.RetinaNetTask(params.task)
    elif isinstance(params.task, configs.maskrcnn.MaskRCNNTask):
      task = tasks.maskrcnn.MaskRCNNTask(params.task)
    elif isinstance(params.task,
                    configs.semantic_segmentation.SemanticSegmentationTask):
      task = tasks.semantic_segmentation.SemanticSegmentationTask(params.task)
    else:
      raise ValueError('Task {} not supported.'.format(type(params.task)))
  # Ensure batch size is 1 for TFLite model.
  params.task.train_data.global_batch_size = 1
  params.task.train_data.dtype = 'float32'
  logging.info('Task config: %s', params.task.as_dict())
  return task.build_inputs(params=params.task.train_data)


def representative_dataset(
    params: cfg.ExperimentConfig,
    task: Optional[base_task.Task] = None,
    calibration_steps: int = 2000) -> Iterator[List[tf.Tensor]]:
  """"Creates representative dataset for input calibration.

  Args:
    params: An ExperimentConfig.
    task: An optional task instance. If it is None, task will be built according
      to the task type in params.
    calibration_steps: The steps to do calibration.

View on GitHub (pinned to e006f5f0d5)

Solutions

1. Use a supported task for TFLite export (e.g. 'classification', 'detection', 'segmentation').
2. Check the task name passed to the TFLite export library.

When it happens

Trigger: Thrown at official/vision/serving/export_tflite_lib.py:57 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/a08def36a3c2747b. Report an issue: GitHub.