@radekosmulski thanks a lot !

I think there could be something to do with docs as failing job here is real :

• https://github.com/pytorch/ignite/pull/1982/checks?check_run_id=2482536753

Could you also please put "Fixes #1939" instead of "Implements" to close the issue automatically once the PR is merged. Thanks

@radekosmulski can we improve the docstring example as well :

eos = EpochOutputStore()
trainer = create_supervised_trainer(model, optimizer, loss)
train_evaluator = create_supervised_evaluator(model, metrics)
eos.attach(train_evaluator, 'output')

@trainer.on(Events.EPOCH_COMPLETED)
def log_training_results(engine):
    train_evaluator.run(train_loader)
    output = train_evaluator.output
    # output = [(y_pred0, y0), (y_pred1, y1), ...] 
    # do something with output, e.g., plotting

It would be also very helpful to add Attributes field and describe data as list of engine outputs, optionally transformed by output_transform.
For more details on google style: https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html

I also think we should fix typing hint here:

- self.data = []  # type: List[Union[int, Tuple[int, int]]]
+ self.data = []  # type: List[Any]