pytorch - 💡(How to fix) Fix DISABLED test_update_batch_norm_mechanism_make_functional_originally_track_running_stats_False_cpu (__main__.TestExamplesCorrectnessCPU) [1 comments, 1 participants]

Traceback (most recent call last): File "/var/lib/jenkins/workspace/test/functorch/test_eager_transforms.py", line 4471, in test_update_batch_norm nn.Conv2d(64, 64, 3), File "/var/lib/jenkins/workspace/test/functorch/test_eager_transforms.py", line 4472, in torch_dynamo_resume_in_test_update_batch_norm_at_4471 nn.BatchNorm2d( File "/var/lib/jenkins/workspace/test/functorch/test_eager_transforms.py", line 4477, in torch_dynamo_resume_in_test_update_batch_norm_at_4472 nn.Linear(43264, classes), File "/var/lib/jenkins/workspace/test/functorch/test_eager_transforms.py", line 4508, in torch_dynamo_resume_in_test_update_batch_norm_at_4477 torch.autograd.grad(compute_loss(x[i], y[i], params, buffers), flat_params) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ for i in range(num_batches) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ ] ^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/apis.py", line 220, in wrapped return vmap_impl( # pyrefly: ignore[bad-argument-type] ...<6 lines>... **kwargs, ) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/vmap.py", line 316, in vmap_impl return _flat_vmap( func, ...<6 lines>... **kwargs, ) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/vmap.py", line 507, in _flat_vmap batched_outputs = func(*batched_inputs, **kwargs) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/apis.py", line 433, in wrapper return eager_transforms.grad_impl(func, argnums, has_aux, args, kwargs) ~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/eager_transforms.py", line 1533, in grad_impl results = grad_and_value_impl(func, argnums, has_aux, args, kwargs) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/vmap.py", line 56, in fn return f(*args, **kwargs) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/eager_transforms.py", line 1484, in grad_and_value_impl output = func(*args, **kwargs) File "/var/lib/jenkins/workspace/test/functorch/test_eager_transforms.py", line 4491, in compute_loss return criterion(fnet(params, buffers, x), y) ~~~~^^^^^^^^^^^^^^^^^^^^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/module.py", line 1778, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/module.py", line 1789, in _call_impl return forward_call(*args, **kwargs) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/_functorch/make_functional.py", line 331, in forward return self.stateless_model(*args, **kwargs) ~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/module.py", line 1778, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/module.py", line 1789, in _call_impl return forward_call(*args, **kwargs) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/container.py", line 253, in forward input = module(input) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/module.py", line 1778, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/module.py", line 1789, in _call_impl return forward_call(*args, **kwargs) File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/conv.py", line 565, in forward return self._conv_forward(input, self.weight, self.bias) ~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/opt/conda/envs/py_3.14/lib/python3.14t/site-packages/torch/nn/modules/conv.py", line 560, in _conv_forward return F.conv2d( ~~~~~~~~^ input, weight, bias, self.stride, self.padding, self.dilation, self.groups ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ) ^ NotImplementedError: Cannot access storage of TensorWrapper

To execute this test, run the following from the base repo dir: PYTORCH_TEST_WITH_DYNAMO=1 python test/functorch/test_eager_transforms.py TestExamplesCorrectnessCPU.test_update_batch_norm_mechanism_make_functional_originally_track_running_stats_False_cpu

This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TL;DR

The most likely fix for the failing test is to update the torch and functorch libraries to the latest versions and re-run the test.

Guidance

• Investigate the NotImplementedError: Cannot access storage of TensorWrapper exception in the stacktrace to understand the root cause of the issue.
• Review the functorch/test_eager_transforms.py test file and the TestExamplesCorrectnessCPU.test_update_batch_norm_mechanism_make_functional_originally_track_running_stats_False_cpu test case to identify potential issues.
• Run the test with the PYTORCH_TEST_WITH_DYNAMO=1 environment variable set to enable dynamo testing and reproduce the issue.
• Check the recent examples and workflow logs for more information on the test failures.

Example

No code example is provided as the issue is related to a specific test case and library versions.

Notes

The issue is related to a flaky test, and the provided stacktrace indicates a NotImplementedError exception. The test is disabled, and the CI is green, but the issue is still present. The provided debugging instructions can help identify the root cause of the issue.

Recommendation

Apply a workaround by updating the torch and functorch libraries to the latest versions and re-running the test. This may resolve the issue, but further investigation is required to determine the root cause.