Apply TorchFix TOR203 fixes (#143691)

Codemodded via `torchfix . --select=TOR203 --fix`.
This is a step to unblock https://github.com/pytorch/pytorch/pull/141076
Pull Request resolved: https://github.com/pytorch/pytorch/pull/143691
Approved by: https://github.com/malfet

benchmarks/profiler_benchmark/resnet_memory_profiler.py

@@ -1,4 +1,4 @@
-import torchvision.models as models
+from torchvision import models
 
 import torch
 import torch.autograd.profiler as profiler

functorch/benchmarks/per_sample_grads.py

@@ -1,8 +1,8 @@
 import time
 
-import torchvision.models as models
 from opacus import PrivacyEngine
 from opacus.utils.module_modification import convert_batchnorm_modules
+from torchvision import models
 
 import torch
 import torch.nn as nn

functorch/examples/dp_cifar10/cifar10_opacus.py

@@ -12,9 +12,8 @@ import sys
 from datetime import datetime, timedelta
 
 import numpy as np
-import torchvision.transforms as transforms
 from opacus import PrivacyEngine
-from torchvision import models
+from torchvision import models, transforms
 from torchvision.datasets import CIFAR10
 from tqdm import tqdm
 

functorch/examples/dp_cifar10/cifar10_transforms.py

@@ -12,8 +12,7 @@ import sys
 from datetime import datetime, timedelta
 
 import numpy as np
-import torchvision.transforms as transforms
-from torchvision import models
+from torchvision import models, transforms
 from torchvision.datasets import CIFAR10
 from tqdm import tqdm
 

functorch/examples/maml_omniglot/support/omniglot_loaders.py

@@ -22,8 +22,8 @@ import os
 import os.path
 
 import numpy as np
-import torchvision.transforms as transforms
 from PIL import Image
+from torchvision import transforms
 
 import torch
 import torch.utils.data as data

test/functorch/test_eager_transforms.py

@@ -4541,7 +4541,7 @@ class TestExamplesCorrectness(TestCase):
     @unittest.skipIf(not USE_TORCHVISION, "test requires torchvision")
     @parametrize("mechanism", ["make_functional", "functional_call"])
     def test_resnet18_per_sample_grads(self, device, mechanism):
-        import torchvision.models as models
+        from torchvision import models
 
         model = models.__dict__["resnet18"](
             pretrained=False, norm_layer=(lambda c: nn.GroupNorm(min(32, c), c))