doctest fix

  ...     sklearn.metrics.mean_absolute_error)

  ...     sklearn.metrics.mean_absolute_error)

  Evaluators can also be used with `dc.metrics.Metric` objects as well

  Evaluators can also be used with `dc.metrics.Metric` objects as well

  in case you want to customize your metric further. 

  in case you want to customize your metric further. (Note that a given

  generator can only be used once so we have to redefine the generator here.)

  >>> generator = model.default_generator(dataset, pad_batches=False)

  >>> evaluator = GeneratorEvaluator(model, generator, transformers)

  >>> metric = dc.metrics.Metric(dc.metrics.mae_score)

  >>> metric = dc.metrics.Metric(dc.metrics.mae_score)

  >>> multitask_scores = evaluator.compute_model_performance(metric)

  >>> multitask_scores = evaluator.compute_model_performance(metric)

  """

  """