Revert PR Lightning-AI#9895

Author: ninginthecloud

CHANGELOG.md

@@ -186,8 +186,11 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 - Added support for `torch.autograd.set_detect_anomaly` through `Trainer` constructor argument `detect_anomaly` ([#9848](https://github.com/PyTorchLightning/pytorch-lightning/pull/9848))
 
+<<<<<<< HEAD
 
 - Added a `len` method to `LightningDataModule` ([#9895](https://github.com/PyTorchLightning/pytorch-lightning/pull/9895))
+=======
+>>>>>>> parent of 6429de894 (Add support for `len(datamodule)` (#9895))
 
 
 - Added `enable_model_summary` flag to Trainer ([#9699](https://github.com/PyTorchLightning/pytorch-lightning/pull/9699))

pytorch_lightning/core/datamodule.py

@@ -22,10 +22,7 @@
 from pytorch_lightning.core.hooks import CheckpointHooks, DataHooks
 from pytorch_lightning.core.mixins import HyperparametersMixin
 from pytorch_lightning.utilities import rank_zero_deprecation
-from pytorch_lightning.utilities.apply_func import apply_to_collection
 from pytorch_lightning.utilities.argparse import add_argparse_args, from_argparse_args, get_init_arguments_and_types
-from pytorch_lightning.utilities.data import has_len
-from pytorch_lightning.utilities.warnings import rank_zero_warn
 
 
 class LightningDataModule(CheckpointHooks, DataHooks, HyperparametersMixin):
@@ -484,6 +481,7 @@ def __getstate__(self) -> dict:
         for fn in ("prepare_data", "setup", "teardown"):
             del d[fn]
         return d
+<<<<<<< HEAD
 
     def __len__(self) -> int:
         """Returns the total number of batches in all dataloaders defined in the datamodule."""
@@ -521,3 +519,5 @@ def get_num_batches(dataloader: DataLoader, name: str) -> None:
             rank_zero_warn("You datamodule does not have any valid dataloader so `__len__` will be returned as 0.")
 
         return num_batches
+=======
+>>>>>>> parent of 6429de894 (Add support for `len(datamodule)` (#9895))

tests/core/test_datamodules.py

@@ -21,15 +21,13 @@
 import pytest
 import torch
 from omegaconf import OmegaConf
-from torch.utils.data import DataLoader
 
 from pytorch_lightning import LightningDataModule, Trainer
 from pytorch_lightning.callbacks import ModelCheckpoint
-from pytorch_lightning.trainer.supporters import CombinedLoader
 from pytorch_lightning.utilities import AttributeDict
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
 from pytorch_lightning.utilities.model_helpers import is_overridden
-from tests.helpers import BoringDataModule, BoringModel, RandomDataset
+from tests.helpers import BoringDataModule, BoringModel
 from tests.helpers.datamodules import ClassifDataModule
 from tests.helpers.runif import RunIf
 from tests.helpers.simple_models import ClassificationModel
@@ -566,14 +564,13 @@ class BoringDataModule1(LightningDataModule):
         batch_size: int
         dims: int = 2
 
-        def train_dataloader(self):
-            return DataLoader(torch.randn(self.batch_size * 2, 10), batch_size=self.batch_size)
+        def __post_init__(self):
+            super().__init__(dims=self.dims)
 
     # asserts for the different dunder methods added by dataclass, when __init__ is implemented, i.e.
     # __repr__, __eq__, __lt__, __le__, etc.
     assert BoringDataModule1(batch_size=64).dims == 2
     assert BoringDataModule1(batch_size=32)
-    assert len(BoringDataModule1(batch_size=32)) == 2
     assert hasattr(BoringDataModule1, "__repr__")
     assert BoringDataModule1(batch_size=32) == BoringDataModule1(batch_size=32)
 
@@ -584,9 +581,7 @@ class BoringDataModule2(LightningDataModule):
 
     # asserts for the different dunder methods added by dataclass, when super class is inherently initialized, i.e.
     # __init__, __repr__, __eq__, __lt__, __le__, etc.
-    assert BoringDataModule2(batch_size=32) is not None
-    assert BoringDataModule2(batch_size=32).batch_size == 32
-    assert len(BoringDataModule2(batch_size=32)) == 0
+    assert BoringDataModule2(batch_size=32)
     assert hasattr(BoringDataModule2, "__repr__")
     assert BoringDataModule2(batch_size=32).prepare_data() is None
     assert BoringDataModule2(batch_size=32) == BoringDataModule2(batch_size=32)