Issue with DataLoader with _dataset_kind = Iterable

[vfdev-5]

The following code wont work:

import torch

class MyIterableDataset(torch.utils.data.IterableDataset):
    def __init__(self, start, end):
        super(MyIterableDataset).__init__()
        assert end > start, "this example code only works with end >= start"
        self.start = start
        self.end = end

    def __iter__(self):
        return iter(range(self.start, self.end))

ds = MyIterableDataset(0, 1000)
data_loader = torch.utils.data.DataLoader(ds, num_workers=2)

from ignite.engine import Engine

def foo(e, b):
    print("{}-{}: {}".format(e.state.epoch, e.state.iteration, b))

engine = Engine(foo)
engine.run(data_loader, epoch_length=10)

and gives the error

ValueErrorTraceback (most recent call last)
<ipython-input-19-1c8004fbf46e> in <module>
     21 
     22 engine = Engine(foo)
---> 23 engine.run(data_loader, epoch_length=10)

/opt/conda/lib/python3.7/site-packages/ignite/engine/engine.py in run(self, data, max_epochs, epoch_length, seed)
    848 
    849         self.state.dataloader = data
--> 850         return self._internal_run()
    851 
    852     def _setup_engine(self):

/opt/conda/lib/python3.7/site-packages/ignite/engine/engine.py in _internal_run(self)
    950             self._dataloader_iter = self._dataloader_len = None
    951             self.logger.error("Engine run is terminating due to exception: %s.", str(e))
--> 952             self._handle_exception(e)
    953 
    954         self._dataloader_iter = self._dataloader_len = None

/opt/conda/lib/python3.7/site-packages/ignite/engine/engine.py in _handle_exception(self, e)
    714             self._fire_event(Events.EXCEPTION_RAISED, e)
    715         else:
--> 716             raise e
    717 
    718     def state_dict(self):

/opt/conda/lib/python3.7/site-packages/ignite/engine/engine.py in _internal_run(self)
    933 
    934                 if self._dataloader_iter is None:
--> 935                     self._setup_engine()
    936 
    937                 hours, mins, secs = self._run_once_on_dataset()

/opt/conda/lib/python3.7/site-packages/ignite/engine/engine.py in _setup_engine(self)
    873             if not isinstance(batch_sampler, ReproducibleBatchSampler):
    874                 self.state.dataloader = _update_dataloader(self.state.dataloader,
--> 875                                                            ReproducibleBatchSampler(batch_sampler))
    876 
    877         iteration = self.state.iteration

/opt/conda/lib/python3.7/site-packages/ignite/engine/engine.py in _update_dataloader(dataloader, new_batch_sampler)
    963     params = {k: getattr(dataloader, k) for k in params_keys}
    964     params['batch_sampler'] = new_batch_sampler
--> 965     return torch.utils.data.DataLoader(**params)
    966 
    967 

/opt/conda/lib/python3.7/site-packages/torch/utils/data/dataloader.py in __init__(self, dataset, batch_size, shuffle, sampler, batch_sampler, num_workers, collate_fn, pin_memory, drop_last, timeout, worker_init_fn, multiprocessing_context)
    182                 raise ValueError(
    183                     "DataLoader with IterableDataset: expected unspecified "
--> 184                     "batch_sampler option, but got batch_sampler={}".format(batch_sampler))
    185         else:
    186             self._dataset_kind = _DatasetKind.Map

ValueError: DataLoader with IterableDataset: expected unspecified batch_sampler option, but got batch_sampler=<ignite.engine.engine.ReproducibleBatchSampler object at 0x7f6000e3fad0>

Patching it like here can be a workaround:

engine.run(map(lambda x: x, data_loader), epoch_length=10)