Fix IterableDataset with __len__ in Trainer (#8095)

Fix IterableDataset with len in Trainer (#8095)
286dc19a · Jonathan Chang · GitHub · d93acd6f · 286dc19a
Unverified Commit 286dc19a authored Oct 27, 2020 by Jonathan Chang Committed by GitHub Oct 27, 2020
Show whitespace changes
Inline Side-by-side

Showing with 3 additions and 1 deletion

src/transformers/trainer.py src/transformers/trainer.py +3 -1

No files found.
--- a/src/transformers/trainer.py
+++ b/src/transformers/trainer.py
@@ -384,7 +384,9 @@ class Trainer:
        dataset.set_format(type=dataset.format["type"], columns=columns)

    def _get_train_sampler(self) -> Optional[torch.utils.data.sampler.Sampler]:
-        if not isinstance(self.train_dataset, collections.abc.Sized):
+        if isinstance(self.train_dataset, torch.utils.data.IterableDataset) or not isinstance(
+            self.train_dataset, collections.abc.Sized
+        ):
            return None
        elif is_torch_tpu_available():
            return get_tpu_sampler(self.train_dataset)