tensorflow/models · error · ValueError

The tasks argument has an invalid type: %s

Error message

The tasks argument has an invalid type: %s

What it means

Error "The tasks argument has an invalid type: %s" thrown in tensorflow/models.

Source

Thrown at official/modeling/multitask/multitask.py:62

      tasks: a list or a flat dict of Task.
      task_weights: a dict of (task, task weight), task weight can be applied
        directly during loss summation in a joint backward step, or it can be
        used to sample task among interleaved backward step.
      task_eval_steps: a dict of (task, eval steps).
      name: the instance name of a MultiTask object.
    """
    super().__init__(name=name)
    if isinstance(tasks, list):
      self._tasks = {}
      for task in tasks:
        if task.name in self._tasks:
          raise ValueError("Duplicated tasks found, task.name is %s" %
                           task.name)
        self._tasks[task.name] = task
    elif isinstance(tasks, dict):
      self._tasks = tasks
    else:
      raise ValueError("The tasks argument has an invalid type: %s" %
                       type(tasks))
    self.task_eval_steps = task_eval_steps or {}
    self._task_weights = task_weights or {}
    self._task_weights = dict([
        (name, self._task_weights.get(name, 1.0)) for name in self.tasks
    ])

  @classmethod
  def from_config(cls, config: configs.MultiTaskConfig, logging_dir=None):
    tasks = {}
    task_eval_steps = {}
    task_weights = {}
    for task_routine in config.task_routines:
      task_name = task_routine.task_name or task_routine.task_config.name
      tasks[task_name] = task_factory.get_task(
          task_routine.task_config, logging_dir=logging_dir, name=task_name)
      task_eval_steps[task_name] = task_routine.eval_steps
      task_weights[task_name] = task_routine.task_weight

View on GitHub (pinned to e006f5f0d5)

When it happens

Trigger: Thrown at official/modeling/multitask/multitask.py:62 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/8779afc08096973d. Report an issue: GitHub.