pytorch - 💡(How to fix) Fix DISABLED test_hessian_argnums (__main__.FuncTorchHigherOrderOpTestsWithCompiledAutograd)

Traceback (most recent call last): File "/var/lib/jenkins/workspace/test/dynamo/test_higher_order_ops.py", line 3929, in test_hessian_argnums wrapped_gm = self._compile_check(wrapper_fn, (x, y)) File "/var/lib/jenkins/workspace/test/dynamo/test_higher_order_ops.py", line 3789, in _compile_check expected = torch.compile(fn, backend=backend, fullgraph=fullgraph)(*inputs) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_dynamo/eval_frame.py", line 1051, in compile_wrapper raise e.with_traceback(None) from e.cause # User compiler error torch._dynamo.exc.Unsupported: Attempted to call a super() attribute that is not a function or method Explanation: Dynamo does not know how to trace the call super().unflatten() because super().unflatten is not a function or method attribute. Hint: Ensure the attribute accessed via super() is a standard method or function.

Developer debug context: call_method SuperVariable() unflatten

For more details about this graph break, please visit: https://meta-pytorch.github.io/compile-graph-break-site/gb/gb0006.html

from user code: File "/var/lib/jenkins/workspace/test/dynamo/test_higher_order_ops.py", line 3925, in wrapper_fn return torch.func.hessian(fn, argnums=(1,))(x, y) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_functorch/eager_transforms.py", line 1401, in wrapper_fn jac_outs_ins = tuple( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_functorch/eager_transforms.py", line 1402, in <genexpr> tuple( File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_functorch/eager_transforms.py", line 1403, in <genexpr> safe_unflatten(jac_out_in, -1, primal.shape) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_functorch/eager_transforms.py", line 1247, in safe_unflatten return tensor.unflatten(dim, shape) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/utils/_device.py", line 122, in torch_function return func(*args, **kwargs) File "/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/_tensor.py", line 1456, in unflatten return super().unflatten(dim, sizes)

Set TORCHDYNAMO_VERBOSE=1 for the internal stack trace (please do this especially if you're reporting a bug to PyTorch). For even more developer context, set TORCH_LOGS="+dynamo"

To execute this test, run the following from the base repo dir: PYTORCH_TEST_CUDA_MEM_LEAK_CHECK=1 PYTORCH_TEST_WITH_SLOW_GRADCHECK=1 python test/dynamo/test_higher_order_ops.py FuncTorchHigherOrderOpTestsWithCompiledAutograd.test_hessian_argnums

This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TL;DR

The test test_hessian_argnums is failing due to an unsupported operation in Dynamo, and setting TORCHDYNAMO_VERBOSE=1 may provide more insight into the issue.

Guidance

• The error message indicates that Dynamo does not know how to trace the call super().unflatten() because super().unflatten is not a function or method attribute. This suggests that the issue may be related to the implementation of the unflatten method.
• To debug the issue, set TORCHDYNAMO_VERBOSE=1 to get the internal stack trace, and consider setting TORCH_LOGS="+dynamo" for more developer context.
• The test can be executed manually by running PYTORCH_TEST_CUDA_MEM_LEAK_CHECK=1 PYTORCH_TEST_WITH_SLOW_GRADCHECK=1 python test/dynamo/test_higher_order_ops.py FuncTorchHigherOrderOpTestsWithCompiledAutograd.test_hessian_argnums from the base repo dir.
• Reviewing the logs and the implementation of the unflatten method may provide more insight into the issue.

Example

No code snippet is provided as the issue is related to a specific implementation detail that is not clearly described in the issue.

Notes

The issue may be related to a limitation or bug in Dynamo, and setting the verbose mode may provide more information about the issue. However, without more context or information about the implementation, it is difficult to provide a more specific solution.

Recommendation

Apply workaround: Set TORCHDYNAMO_VERBOSE=1 and review the logs to gain more insight into the issue, as this may provide more information about the unsupported operation and help identify a potential solution.