[bugfix] show_fix should set num_workers to 0

.gitignore

@@ -148,3 +148,4 @@ imdb
 xsum
 coco128
 wmt_en_ro
+kinetics

flash/data/data_module.py

@@ -21,7 +21,7 @@
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
 from torch.nn import Module
 from torch.utils.data import DataLoader, Dataset
-from torch.utils.data.dataset import Subset
+from torch.utils.data.dataset import IterableDataset, Subset
 
 from flash.data.auto_dataset import AutoDataset
 from flash.data.base_viz import BaseVisualization
@@ -138,8 +138,12 @@ def data_fetcher(self, data_fetcher: BaseDataFetcher) -> None:
 
     def _reset_iterator(self, stage: RunningStage) -> Iterable[Any]:
         iter_name = f"_{stage}_iter"
+        # the num_workers has to be set to 0 to work properly.
+        num_workers = self.num_workers
+        self.num_workers = 0
         dataloader_fn = getattr(self, f"{stage}_dataloader")
         iterator = iter(dataloader_fn())
+        self.num_workers = num_workers
         setattr(self, iter_name, iterator)
         return iterator
 
@@ -191,7 +195,8 @@ def get_dataset_attribute(dataset: torch.utils.data.Dataset, attr_name: str, def
     def set_dataset_attribute(dataset: torch.utils.data.Dataset, attr_name: str, value: Any) -> None:
         if isinstance(dataset, Subset):
             dataset = dataset.dataset
-        setattr(dataset, attr_name, value)
+        if isinstance(dataset, (Dataset, IterableDataset)):
+            setattr(dataset, attr_name, value)
 
     def set_running_stages(self):
         if self._train_ds:

tests/data/test_callbacks.py

@@ -181,3 +181,15 @@ def extract_data(data):
         assert dm.data_fetcher.show_collate_called
         assert dm.data_fetcher.per_batch_transform_called
         dm.data_fetcher.check_reset()
+
+
+def test_data_loaders_num_workers_to_0(tmpdir):
+    """
+    The num_workers should be set to `0` internally for visualization and not for training.
+    """
+
+    datamodule = DataModule(train_dataset=range(10), num_workers=3)
+    iterator = datamodule._reset_iterator(RunningStage.TRAINING)
+    assert isinstance(iterator, torch.utils.data.dataloader._SingleProcessDataLoaderIter)
+    iterator = iter(datamodule.train_dataloader())
+    assert isinstance(iterator, torch.utils.data.dataloader._MultiProcessingDataLoaderIter)