diff --git a/torch/testing/_internal/common_modules.py b/torch/testing/_internal/common_modules.py
index ffd0e6f95a87d..9ae062df85ae8 100644
--- a/torch/testing/_internal/common_modules.py
+++ b/torch/testing/_internal/common_modules.py
@@ -25,7 +25,7 @@
     nllloss_reference, nlllossNd_reference, smoothl1loss_reference, softmarginloss_reference, get_reduction)
 from torch.testing._internal.common_utils import (
     freeze_rng_state, set_single_threaded_if_parallel_tbb, skipIfMps, GRADCHECK_NONDET_TOL, TEST_WITH_ROCM, IS_WINDOWS,
-    skipIfTorchDynamo)
+    skipIfTorchDynamo, TEST_WITH_TORCHINDUCTOR)
 from types import ModuleType
 from typing import List, Tuple, Type, Set, Dict
 import operator
@@ -127,7 +127,7 @@ def _parametrize_test(self, test, generic_cls, device_cls):
                     def test_wrapper(*args, **kwargs):
                         return test(*args, **kwargs)
 
-                    if self.skip_if_dynamo and not torch.testing._internal.common_utils.TEST_WITH_TORCHINDUCTOR:
+                    if self.skip_if_dynamo and not TEST_WITH_TORCHINDUCTOR:
                         test_wrapper = skipIfTorchDynamo("Policy: we don't run ModuleInfo tests w/ Dynamo")(test_wrapper)
 
                     decorator_fn = partial(module_info.get_decorators, generic_cls.__name__,
@@ -3469,7 +3469,12 @@ def module_error_inputs_torch_nn_Pad3d(module_info, device, dtype, requires_grad
                        unittest.expectedFailure, 'TestEagerFusionModuleInfo',
                        'test_aot_autograd_module_exhaustive',
                        active_if=operator.itemgetter('training')
-                   ),)
+                   ),
+                   # test fails if run alone in inductor https://github.com/pytorch/pytorch/issues/125967
+                   DecorateInfo(
+                       unittest.skip("Skipped https://github.com/pytorch/pytorch/issues/125967"),
+                       'TestModule', 'test_memory_format', device_type='cuda',
+                       active_if=(TEST_WITH_TORCHINDUCTOR)),)
                ),
     ModuleInfo(torch.nn.BatchNorm3d,
                train_and_eval_differ=True,