[misc] update pre-commit and run all files (#4752)

* [misc] update pre-commit

* [misc] run pre-commit

* [misc] remove useless configuration files

* [misc] ignore cuda for clang-format

tests/test_legacy/test_data/test_cifar10_dataset.py

@@ -4,7 +4,6 @@
 import os
 from pathlib import Path
 
-import pytest
 from torch.utils.data import DataLoader
 from torchvision import datasets, transforms
 
@@ -15,7 +14,7 @@ def test_cifar10_dataset():
     transform_pipeline = transforms.Compose(transform_pipeline)
 
     # build dataset
-    dataset = datasets.CIFAR10(root=Path(os.environ['DATA']), train=True, download=True, transform=transform_pipeline)
+    dataset = datasets.CIFAR10(root=Path(os.environ["DATA"]), train=True, download=True, transform=transform_pipeline)
 
     # build dataloader
     dataloader = DataLoader(dataset=dataset, batch_size=4, shuffle=True, num_workers=2)
@@ -23,5 +22,5 @@ def test_cifar10_dataset():
     img, label = data_iter.next()
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     test_cifar10_dataset()

tests/test_legacy/test_data/test_data_parallel_sampler.py

@@ -4,7 +4,6 @@
 import os
 from pathlib import Path
 
-import pytest
 import torch
 import torch.distributed as dist
 from torchvision import datasets, transforms
@@ -16,24 +15,26 @@ from colossalai.legacy.core import global_context as gpc
 from colossalai.legacy.utils import get_dataloader
 from colossalai.testing import rerun_if_address_is_in_use, spawn
 
-CONFIG = Config(dict(
-    parallel=dict(
-        pipeline=dict(size=1),
-        tensor=dict(size=1, mode=None),
-    ),
-    seed=1024,
-))
+CONFIG = Config(
+    dict(
+        parallel=dict(
+            pipeline=dict(size=1),
+            tensor=dict(size=1, mode=None),
+        ),
+        seed=1024,
+    )
+)
 
 
 def run_data_sampler(rank, world_size, port):
-    dist_args = dict(config=CONFIG, rank=rank, world_size=world_size, backend='gloo', port=port, host='localhost')
+    dist_args = dict(config=CONFIG, rank=rank, world_size=world_size, backend="gloo", port=port, host="localhost")
     colossalai.legacy.launch(**dist_args)
-    print('finished initialization')
+    print("finished initialization")
 
     # build dataset
     transform_pipeline = [transforms.ToTensor()]
     transform_pipeline = transforms.Compose(transform_pipeline)
-    dataset = datasets.CIFAR10(root=Path(os.environ['DATA']), train=True, download=True, transform=transform_pipeline)
+    dataset = datasets.CIFAR10(root=Path(os.environ["DATA"]), train=True, download=True, transform=transform_pipeline)
 
     # build dataloader
     dataloader = get_dataloader(dataset, batch_size=8, add_sampler=True)
@@ -50,7 +51,8 @@ def run_data_sampler(rank, world_size, port):
 
     if gpc.get_local_rank(ParallelMode.DATA) != 0:
         assert not torch.equal(
-            img, img_to_compare), 'Same image was distributed across ranks but expected it to be different'
+            img, img_to_compare
+        ), "Same image was distributed across ranks but expected it to be different"
     torch.cuda.empty_cache()
 
 
@@ -59,5 +61,5 @@ def test_data_sampler():
     spawn(run_data_sampler, 4)
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     test_data_sampler()

tests/test_legacy/test_data/test_deterministic_dataloader.py

@@ -4,7 +4,6 @@
 import os
 from pathlib import Path
 
-import pytest
 import torch
 import torch.distributed as dist
 from torchvision import datasets, transforms
@@ -20,8 +19,8 @@ CONFIG = Config(
     dict(
         train_data=dict(
             dataset=dict(
-                type='CIFAR10',
-                root=Path(os.environ['DATA']),
+                type="CIFAR10",
+                root=Path(os.environ["DATA"]),
                 train=True,
                 download=True,
             ),
@@ -32,17 +31,18 @@ CONFIG = Config(
             tensor=dict(size=1, mode=None),
         ),
         seed=1024,
-    ))
+    )
+)
 
 
 def run_data_sampler(rank, world_size, port):
-    dist_args = dict(config=CONFIG, rank=rank, world_size=world_size, backend='gloo', port=port, host='localhost')
+    dist_args = dict(config=CONFIG, rank=rank, world_size=world_size, backend="gloo", port=port, host="localhost")
     colossalai.legacy.launch(**dist_args)
 
     # build dataset
     transform_pipeline = [transforms.ToTensor(), transforms.RandomCrop(size=32, padding=4)]
     transform_pipeline = transforms.Compose(transform_pipeline)
-    dataset = datasets.CIFAR10(root=Path(os.environ['DATA']), train=True, download=True, transform=transform_pipeline)
+    dataset = datasets.CIFAR10(root=Path(os.environ["DATA"]), train=True, download=True, transform=transform_pipeline)
 
     # build dataloader
     dataloader = get_dataloader(dataset, batch_size=8, add_sampler=False)
@@ -60,8 +60,9 @@ def run_data_sampler(rank, world_size, port):
     if gpc.get_local_rank(ParallelMode.DATA) != 0:
         # this is without sampler
         # this should be false if data parallel sampler to given to the dataloader
-        assert torch.equal(img,
-                           img_to_compare), 'Same image was distributed across ranks and expected it to be the same'
+        assert torch.equal(
+            img, img_to_compare
+        ), "Same image was distributed across ranks and expected it to be the same"
     torch.cuda.empty_cache()
 
 
@@ -70,5 +71,5 @@ def test_data_sampler():
     spawn(run_data_sampler, 4)
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     test_data_sampler()