No metrics in `fit()` causes error

[beekill95]

When I use SimpleTrainingStep without setting metrics (metrics default to None) and with both y_true & y_pred are dictionaries, the following error occurs:

  File "lib/python3.8/site-packages/training_loop/training_loops/training_loop.py", line 150, in fit
    logs = step.train_step(model, data, device)
  File "lib/python3.8/site-packages/training_loop/training_loops/simple_training_step.py", line 185, in train_step
    self._update_train_metrics(train_loss=loss, y_pred=y_pred, y_true=y_true)
  File "lib/python3.8/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
  File "lib/python3.8/site-packages/training_loop/training_loops/simple_training_step.py", line 336, in _update_train_metrics
    update_metrics(self._train_metrics, y_pred=y_pred, y_true=y_true)
  File "lib/python3.8/site-packages/training_loop/training_loops/simple_training_step.py", line 626, in update_metrics
    raise ValueError(
ValueError: If `metrics` is a list, then both `y_true` and `y_pred` must either be a single tensor or a sequence of tensors.