Relax restrictions on wrapping a custom batch sampler in predict (#19678

)

src/lightning/pytorch/CHANGELOG.md

@@ -22,7 +22,7 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 - The `prepare_data()` hook in `LightningModule` and `LightningDataModule` is now subject to a barrier without timeout to avoid long-running tasks to be interrupted ([#19448](https://github.com/Lightning-AI/lightning/pull/19448))
 
--
+- Relaxed the requirement for custom batch samplers to expose `drop_last` for prediction ([#19678](https://github.com/Lightning-AI/pytorch-lightning/pull/19678))
 
 -
 

src/lightning/pytorch/utilities/data.py

@@ -28,6 +28,7 @@
     has_iterable_dataset,
     sized_len,
 )
+from lightning.fabric.utilities.warnings import PossibleUserWarning
 from lightning.pytorch.overrides.distributed import _IndexBatchSamplerWrapper
 from lightning.pytorch.trainer.states import RunningStage
 from lightning.pytorch.utilities.exceptions import MisconfigurationException
@@ -301,6 +302,14 @@ def _dataloader_init_kwargs_resolve_sampler(
                     " or set `Trainer(use_distributed_sampler=False)`. If you choose the latter, you will be"
                     " responsible for handling the distributed sampling within your batch sampler."
                 ) from ex
+        elif is_predicting:
+            rank_zero_warn(
+                f"You are using a custom batch sampler `{batch_sampler_cls.__qualname__}` for prediction."
+                " Lightning would normally set `drop_last=False` to ensure all samples are returned, but for"
+                " custom samplers it can't guarantee this. Make sure your sampler is configured correctly to return"
+                " all indices.",
+                category=PossibleUserWarning,
+            )
         else:
             # The sampler is not a PyTorch `BatchSampler`, we don't know how to inject a custom sampler or
             # how to adjust the `drop_last` value

tests/tests_pytorch/utilities/test_data.py

@@ -5,6 +5,7 @@
 import pytest
 import torch
 from lightning.fabric.utilities.data import _replace_dunder_methods
+from lightning.fabric.utilities.warnings import PossibleUserWarning
 from lightning.pytorch import Trainer
 from lightning.pytorch.demos.boring_classes import RandomDataset, RandomIterableDataset
 from lightning.pytorch.overrides.distributed import _IndexBatchSamplerWrapper
@@ -230,7 +231,8 @@ def __len__(self) -> int:
     assert batch_sampler.drop_last == (not predicting)
 
 
-def test_custom_batch_sampler():
+@pytest.mark.parametrize("predicting", [True, False])
+def test_custom_batch_sampler(predicting):
     """Test that a custom (non-PyTorch) batch sampler requires the user to set `use_distributed_sampler=False`."""
 
     class CustomBatchSampler:  # not inheriting from `BatchSampler`
@@ -240,8 +242,13 @@ def __iter__(self):
 
     batch_sampler = CustomBatchSampler()
     dataloader = DataLoader(range(100), batch_sampler=batch_sampler)
-    with pytest.raises(TypeError, match=r"can't inject a \(distributed\) sampler into your batch sampler"):
-        _ = _update_dataloader(dataloader, sampler=Mock())
+
+    if predicting:
+        with pytest.warns(PossibleUserWarning, match=r"Make sure your sampler is configured correctly to return all"):
+            _ = _update_dataloader(dataloader, sampler=Mock(), mode=RunningStage.PREDICTING)
+    else:
+        with pytest.raises(TypeError, match=r"can't inject a \(distributed\) sampler into your batch sampler"):
+            _ = _update_dataloader(dataloader, sampler=Mock(), mode=None)
 
 
 def test_custom_batch_sampler_no_drop_last():